3D spatiotemporally scalable in vivo neural probes based on fluorinated elastomers.

Electronic devices for recording neural activity in the nervous system need to be scalable across large spatial and temporal scales while also providing millisecond and single-cell spatiotemporal resolution. However, existing high-resolution neural recording devices cannot achieve simultaneous scalability on both spatial and temporal levels due to a trade-off between sensor density and mechanical flexibility. Here we introduce a three-dimensional (3D) stacking implantable electronic platform, based on perfluorinated dielectric elastomers and tissue-level soft multilayer electrodes, that enables spatiotemporally scalable single-cell neural electrophysiology in the nervous system. Our elastomers exhibit stable dielectric performance for over a year in physiological solutions and are 10,000 times softer than conventional plastic dielectrics. By leveraging these unique characteristics we develop the packaging of lithographed nanometre-thick electrode arrays in a 3D configuration with a cross-sectional density of 7.6 electrodes per 100 µm2. The resulting 3D integrated multilayer soft electrode array retains tissue-level flexibility, reducing chronic immune responses in mouse neural tissues, and demonstrates the ability to reliably track electrical activity in the mouse brain or spinal cord over months without disrupting animal behaviour.

Tracking neural activity from the same cells during the entire adult life of mice.

Stably recording the electrical activity of the same neurons over the adult life of an animal is important to neuroscience research and biomedical applications. Current implantable devices cannot provide stable recording on this timescale. Here, we introduce a method to precisely implant electronics with an open, unfolded mesh structure across multiple brain regions in the mouse. The open mesh structure forms a stable interwoven structure with the neural network, preventing probe drifting and showing no immune response and neuron loss during the year-long implantation. Rigorous statistical analysis, visual stimulus-dependent measurement and unbiased, machine-learning-based analysis demonstrated that single-unit action potentials have been recorded from the same neurons of behaving mice in a very long-term stable manner. Leveraging this stable structure, we demonstrated that the same neurons can be recorded over the entire adult life of the mouse, revealing the aging-associated evolution of single-neuron activities.

Explainable multi-task learning for multi-modality biological data analysis.

Current biotechnologies can simultaneously measure multiple high-dimensional modalities (e.g., RNA, DNA accessibility, and protein) from the same cells. A combination of different analytical tasks (e.g., multi-modal integration and cross-modal analysis) is required to comprehensively understand such data, inferring how gene regulation drives biological diversity and functions. However, current analytical methods are designed to perform a single task, only providing a partial picture of the multi-modal data. Here, we present UnitedNet, an explainable multi-task deep neural network capable of integrating different tasks to analyze single-cell multi-modality data. Applied to various multi-modality datasets (e.g., Patch-seq, multiome ATAC + gene expression, and spatial transcriptomics), UnitedNet demonstrates similar or better accuracy in multi-modal integration and cross-modal prediction compared with state-of-the-art methods. Moreover, by dissecting the trained UnitedNet with the explainable machine learning algorithm, we can directly quantify the relationship between gene expression and other modalities with cell-type specificity. UnitedNet is a comprehensive end-to-end framework that could be broadly applicable to single-cell multi-modality biology. This framework has the potential to facilitate the discovery of cell-type-specific regulation kinetics across transcriptomics and other modalities.