arXiv subjects

Stephan Quessy

Publications and source records attributed to Stephan Quessy.

3 recordsLinked to original sources

arXiv2026

Automatic Aberration Correction for Transcranial Functional and Super-Resolution Ultrasound Imaging in Rodents and Nonhuman Primates

Skull-induced aberrations remain a major drawback of transcranial ultrasound localization microscopy (ULM), degrading sensitivity and spatial accuracy through microbubble mislocalization, false detections, and imaging artifacts, such as disconnected or duplicated vessels. Here, we present a differentiable beamforming framework for automatic aberration correction in transcranial Doppler and ULM. Our approach uses spatially distributed delay-based parameterization of the aberration that is optimized in a closed-loop manner using angular coherence as an objective function. We demonstrate robust improvements of transcranial ULM, in vivo, with enhanced resolution of both mouse and nonhuman primate (NHP) brains. We also extended differentiable beamforming to functional measurements, with improvements in the sensitivity of transcranial functional ultrasound (fUS) and ULM based hemodynamic quantification. Extending this approach to 3D transcranial ULM imaging in NHPs, we show efficient correction of skull induced aberrations and removal of artifacts, such as vessel duplications. By providing a fully automated and generalizable solution for aberration correction, this work lowers a major technical barrier to transcranial ultrasound imaging, enabling broader adoption of non-invasive, super-resolution and functional neuroimaging across laboratories and across species.

physics.med-ph↗

arXiv2025

NeuroClean: A Generalized Machine-Learning Approach to Neural Time-Series Conditioning

Electroencephalography (EEG) and local field potentials (LFP) are two widely used techniques to record electrical activity from the brain. These signals are used in both the clinical and research domains for multiple applications. However, most brain data recordings suffer from a myriad of artifacts and noise sources other than the brain itself. Thus, a major requirement for their use is proper and, given current volumes of data, a fully automatized conditioning. As a means to this end, here we introduce an unsupervised, multipurpose EEG/LFP preprocessing method, the NeuroClean pipeline. In addition to its completeness and reliability, NeuroClean is an unsupervised series of algorithms intended to mitigate reproducibility issues and biases caused by human intervention. The pipeline is designed as a five-step process, including the common bandpass and line noise filtering, and bad channel rejection. However, it incorporates an efficient independent component analysis with an automatic component rejection based on a clustering algorithm. This machine learning classifier is used to ensure that task-relevant information is preserved after each step of the cleaning process. We used several data sets to validate the pipeline. NeuroClean removed several common types of artifacts from the signal. Moreover, in the context of motor tasks of varying complexity, it yielded more than 97% accuracy (vs. a chance-level of 33.3%) in an optimized Multinomial Logistic Regression model after cleaning the data, compared to the raw data, which performed at 74% accuracy. These results show that NeuroClean is a promising pipeline and workflow that can be applied to future work and studies to achieve better generalization and performance on machine learning pipelines.

cs.LG↗

arXiv2024

Towards Transcranial 3D Ultrasound Localization Microscopy of the Nonhuman Primate Brain

Hemodynamic changes occur in stroke and neurodegenerative diseases. Developing imaging techniques allowing the in vivo visualization and quantification of cerebral blood flow would help better understand the underlying mechanism of those cerebrovascular diseases. 3D ultrasound localization microscopy (ULM) is a novel technology that can map the microvasculature of the brain at large depth and has been mainly used until now in rodents. Here, we demonstrated the feasibility of 3D ULM of the nonhuman primate (NHP) brain with a single 256-channels programmable ultrasound scanner. We achieved a highly resolved vascular map of the macaque brain at large depth in presence of craniotomy and durectomy using an 8-MHz multiplexed matrix probe. We were able to distinguish vessels as small as 26.9 {\mu}m. We also demonstrated that transcranial imaging of the macaque brain at similar depth was feasible using a 3-MHz probe and achieved a resolution of 60.4 {\mu}m. This work paves the way to clinical application of 3D ULM.

eess.IV↗