Exploiting holographically encoded variance to transmit labelled images through a multimode optical fiber.

Artificial intelligence has emerged as promising tool to decode an image transmitted through a multimode fiber (MMF) by applying deep learning techniques. By transmitting thousands of images through the MMF, deep neural networks (DNNs) are able to decipher the seemingly random output speckle patterns and unveil the intrinsic input-output relationship. High fidelity reconstruction is obtained for datasets with a large degree of homogeneity, which underutilizes the capacity of the combined MMF-DNN system. Here, we show that holographic modulation can encode an additional layer of variance on the output speckle pattern, improving the overall transmissive capabilities of the system. Operatively, we have implemented this by adding a holographic label to the original dataset and injecting the resulting phase image into the fiber facet through a Fourier transform lens. The resulting speckle pattern dataset can be clustered primarily by holographic label, and can be reconstructed without loss of fidelity. As an application, we describe how color images may be segmented into RGB components and each color component may then be labelled by distinct hologram. A ResUNet architecture was then used to decode each class of speckle patterns and reconstruct the color image without the need for temporal synchronization between sender and receiver.

Potential of plasmonics and nanoscale light-matter interactions for the next generation of optical neural interfaces.

Within the realm of optical neural interfaces, the exploration of plasmonic resonances to interact with neural cells has captured increasing attention among the neuroscience community. The interplay of light with conduction electrons in nanometer-sized metallic nanostructures can induce plasmonic resonances, showcasing a versatile capability to both sense and trigger cellular events. We describe the perspective of generating propagating or localized surface plasmon polaritons on the tip of an optical neural implant, widening the possibility for neuroscience labs to explore the potential of plasmonic neural interfaces.

Deep autoencoder as an interpretable tool for Raman spectroscopy investigation of chemical and extracellular vesicle mixtures.

Surface-enhanced Raman spectroscopy (SERS) is a powerful tool that provides valuable insight into the molecular contents of chemical and biological samples. However, interpreting Raman spectra from complex or dynamic datasets remains challenging, particularly for highly heterogeneous biological samples like extracellular vesicles (EVs). To overcome this, we developed a tunable and interpretable deep autoencoder for the analysis of several challenging Raman spectroscopy applications, including synthetic datasets, chemical mixtures, a chemical milling reaction, and mixtures of EVs. We compared the results with classical methods (PCA and UMAP) to demonstrate the superior performance of the proposed technique. Our method can handle small datasets, provide a high degree of generalization such that it can fill unknown gaps within spectral datasets, and even quantify relative ratios of cell line-derived EVs to fetal bovine serum-derived EVs within mixtures. This simple yet robust approach will greatly improve the analysis capabilities for many other Raman spectroscopy applications.