Myelinated Fibers Come into Focus with Raman Imaging
Deep learning and coherent Raman imaging provide a label-free view of myelinated fibers in human white matter
A label-free Raman imaging pipeline has provided the first quantitative measurements of axon diameter and myelin thickness in the human uncinate fasciculus, a long-range white matter tract linking the temporal lobe with frontal brain regions involved in decision-making and social behavior.
The approach combines spectral-focusing coherent anti-Stokes Raman spectroscopy (sf-CARS) with a custom AxonDeepSeg deep-learning model. By targeting the CH₂ vibrational signature of lipid-rich myelin, sf-CARS allowed the team to image myelinated fibers without stains or labels, including in postmortem human tissue where electron microscopy is often impractical. The spectral-focusing setup used chirped femtosecond laser pulses to improve spectral resolution and contrast for myelin imaging.
The resulting images were processed with an active-learning segmentation workflow trained to distinguish axons from surrounding myelin sheaths. After quality control, the researchers analyzed 2,645 myelinated axons from six control donors, measuring axon diameter, myelin thickness, and g-ratio.
Axon diameters ranged from 0.37 to 6.38 μm, with a mean of 0.93 μm, while mean myelin thickness was 0.48 μm. A comparison with anterior cingulate cortex white matter from the same individual showed thicker myelin and lower g-ratios in the uncinate fasciculus, consistent with the demands of long-range signaling.
The authors note that the method currently captures only myelinated axons and is sensitive to fiber orientation, but say the pipeline could support larger studies of white matter structure in psychiatric and neurological disorders.
Heavy Metal Aromaticity
A three-atom bismuth ring held between uranium or thorium centers shows all-metal aromaticity in one of the heaviest systems yet confirmed
A three-atom ring made entirely of bismuth has been stabilized between actinide centers, giving chemists a rare example of all-metal aromaticity in one of the heaviest systems yet confirmed. The study, led by researchers at The University of Manchester, centers on diuranium and dithorium “inverse-sandwich” complexes that hold a triangular Bi₃³⁻ unit in place.
Aromaticity is most familiar from organic molecules such as benzene, where delocalized electrons generate stabilizing ring currents. In this case, single-crystal X-ray diffraction confirmed the structure of the bismuth ring, while magnetic measurements, spectroscopy, and computational modeling showed that it can sustain comparable ring currents despite being composed entirely of heavy metal atoms.
The key distinction is that the effect is dominated by sigma electrons, rather than the pi electrons usually associated with organic aromaticity. The dithorium complex also showed measurable exalted diamagnetism, providing experimental support for aromatic ring-current behavior.
“Aromaticity is often taught through benzene, but here we’ve shown a three-atom ring of bismuth – supported by uranium or thorium – can sustain robust, measurable ring currents,” said Steve Liddle in the team’s press release.
“It’s a powerful reminder that the deepest principles of chemical bonding apply far beyond carbon.”
Label-Free Sensing for Viral Vector Production
The workflow combines antibody-coated gold electrodes, impedance spectroscopy, and machine learning for faster AAV2 sensing
Counting viral vectors during gene therapy manufacturing could become faster with a label-free sensing workflow that combines electrochemical impedance spectroscopy (EIS) with machine learning. Developed at North Carolina State University, the platform targets adeno-associated virus serotype 2 (AAV2), a common model for viral vector production.
The sensor uses gold microelectrodes functionalized with anti-AAV2 antibodies. When viral particles bind to the electrode surface, they change the impedance response, but those signals overlap with background variation from pH and nonspecific binding. Instead of relying on equivalent circuit modeling, the team trained machine learning models on impedance-derived features to separate vector-associated signals from process noise.
The system was tested across four pH conditions and AAV2 titers spanning 10⁸ to 10¹² capsids per mL. Logistic regression and k-nearest neighbors classified pH with 94–95 percent test accuracy, while XGBoost performed best for viral titer prediction, reaching a test R² of 0.78 in buffered samples and 0.76 in clarified cell lysate.
“High-speed and rapid sensing is not just for the sake of speed alone,” said co-corresponding author Stefano Menegatti in the team’s press release. “It allows us to implement rapid corrective measures as needed in the processes that produce and purify viral vectors.”
“Ultimately, we think our new approach serves as an extremely useful tool for complementing ELISA in a wide range of contexts – particularly for biomanufacturing,” added co-corresponding author Michael Daniele.
A Better Diagnosis for Electrolyzer Degradation
New method disentangles component-level degradation in AEM electrolyzers without relying on predefined equivalent-circuit models
Researchers at the Korea Institute of Materials Science (KIMS) have developed an in situ diagnostic framework for tracking performance degradation in anion exchange membrane water electrolysis (AEMWE) systems under operating conditions. The approach is designed to address a practical problem in green hydrogen production: conventional two-electrode measurements show overall voltage loss but struggle to identify which cell components are responsible.
The team combined galvanostatic electrochemical impedance spectroscopy (GEIS) with distribution of relaxation times (DRT) analysis, allowing overlapping impedance features to be separated without relying on predefined equivalent-circuit models. Using electrolyte modulation, anode catalyst replacement, and membrane variation, they assigned individual DRT peaks to cathode and anode hydroxide-transfer resistance, charge-transfer resistance, and mass transport processes.
Long-term testing showed that degradation was driven mainly by the cobalt oxide anode, while the platinum/carbon cathode remained comparatively stable. Reconstructed overpotential contributions indicated that performance loss arose primarily from increased anode charge-transfer and hydroxide-transfer resistances, supported by complementary X-ray photoelectron spectroscopy, X-ray diffraction, and microscopy data.
“This study presents a new analytical framework that enables real-time deconvolution and interpretation of voltage loss mechanisms in complex water electrolysis systems under actual operating conditions,” said Sung Mook Choi, lead author of the study and principal researcher at KIMS.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
