A deep learning framework for classifying microglia activation state using morphology and intrinsic fluorescence lifetime data.

Microglia are the immune cell in the central nervous system (CNS) and exist in a surveillant state characterized by a ramified form in the healthy brain. In response to brain injury or disease including neurodegenerative diseases, they become activated and change their morphology. Due to known correlation between this activation and neuroinflammation, there is great interest in improved approaches for studying microglial activation in the context of CNS disease mechanisms. One classic approach has utilized Microglia's morphology as one of the key indicators of its activation and correlated with its functional state. More recently microglial activation has been shown to have intrinsic NADH metabolic signatures that are detectable via fluorescence lifetime imaging (FLIM). Despite the promise of morphology and metabolism as key fingerprints of microglial function, they has not been analyzed together due to lack of an appropriate computational framework. Here we present a deep neural network to study the effect of both morphology and FLIM metabolic signatures toward identifying its activation status. Our model is tested on 1, 000+ cells (ground truth generated using LPS treatment) and provides a state-of-the-art framework to identify microglial activation and its role in neurodegenerative diseases.

Joint regression-classification deep learning framework for analyzing fluorescence lifetime images using NADH and FAD.

In this paper, we develop a deep neural network based joint classification-regression approach to identify microglia, a resident central nervous system macrophage, in the brain using fluorescence lifetime imaging microscopy (FLIM) data. Microglia are responsible for several key aspects of brain development and neurodegenerative diseases. Accurate detection of microglia is key to understanding their role and function in the CNS, and has been studied extensively in recent years. In this paper, we propose a joint classification-regression scheme that can incorporate fluorescence lifetime data from two different autofluorescent metabolic co-enzymes, FAD and NADH, in the same model. This approach not only represents the lifetime data more accurately but also provides the classification engine a more diverse data source. Furthermore, the two components of model can be trained jointly which combines the strengths of the regression and classification methods. We demonstrate the efficacy of our method using datasets generated using mouse brain tissue which show that our joint learning model outperforms results on the coenzymes taken independently, providing an efficient way to classify microglia from other cells.

Rhesus monkeys as a translational model for late-onset Alzheimer's disease.

Age is a major risk factor for late-onset Alzheimer's disease (AD) but seldom features in laboratory models of the disease. Furthermore, heterogeneity in size and density of AD plaques observed in individuals are not recapitulated in transgenic mouse models, presenting an incomplete picture. We show that the amyloid plaque microenvironment is not equivalent between rodent and primate species, and that differences in the impact of AD pathology on local metabolism and inflammation might explain established differences in neurodegeneration and functional decline. Using brain tissue from transgenic APP/PSEN1 mice, rhesus monkeys with age-related amyloid plaques, and human subjects with confirmed AD, we report altered energetics in the plaque microenvironment. Metabolic features included changes in mitochondrial distribution and enzymatic activity, and changes in redox cofactors NAD(P)H that were shared among species. A greater burden of lipofuscin was detected in the brains from monkeys and humans of advanced age compared to transgenic mice. Local inflammatory signatures indexed by astrogliosis and microglial activation were detected in each species; however, the inflamed zone was considerably larger for monkeys and humans. These data demonstrate the advantage of nonhuman primates in modeling the plaque microenvironment, and provide a new framework to investigate how AD pathology might contribute to functional loss.

Machine Learning Methods for Fluorescence Lifetime Imaging (FLIM) Based Label-Free Detection of Microglia.

Automated computational analysis techniques utilizing machine learning have been demonstrated to be able to extract more data from different imaging modalities compared to traditional analysis techniques. One new approach is to use machine learning techniques to existing multiphoton imaging modalities to better interpret intrinsically fluorescent cellular signals to characterize different cell types. Fluorescence Lifetime Imaging Microscopy (FLIM) is a high-resolution quantitative imaging tool that can detect metabolic cellular signatures based on the lifetime variations of intrinsically fluorescent metabolic co-factors such as nicotinamide adenine dinucleotide [NAD(P)H]. NAD(P)H lifetime-based discrimination techniques have previously been used to develop metabolic cell signatures for diverse cell types including immune cells such as macrophages. However, FLIM could be even more effective in characterizing cell types if machine learning was used to classify cells by utilizing FLIM parameters for classification. Here, we demonstrate the potential for FLIM-based, label-free NAD(P)H imaging to distinguish different cell types using Artificial Neural Network (ANN)-based machine learning. For our biological use case, we used the challenge of differentiating microglia from other glia cell types in the brain. Microglia are the resident macrophages of the brain and spinal cord and play a critical role in maintaining the neural environment and responding to injury. Microglia are challenging to identify as most fluorescent labeling approaches cross-react with other immune cell types, are often insensitive to activation state, and require the use of multiple specialized antibody labels. Furthermore, the use of these extrinsic antibody labels prevents application in in vivo animal models and possible future clinical adaptations such as neurodegenerative pathologies. With the ANN-based NAD(P)H FLIM analysis approach, we found that microglia in cell culture mixed with other glial cells can be identified with more than 0.9 True Positive Rate (TPR). We also extended our approach to identify microglia in fixed brain tissue with a TPR of 0.79. In both cases the False Discovery Rate was around 30%. This method can be further extended to potentially study and better understand microglia's role in neurodegenerative disease with improved detection accuracy.

Optical fiber-based dispersion for spectral discrimination in fluorescence lifetime imaging systems.

The excited state lifetime of a fluorophore together with its fluorescence emission spectrum provide information that can yield valuable insights into the nature of a fluorophore and its microenvironment. However, it is difficult to obtain both channels of information in a conventional scheme as detectors are typically configured either for spectral or lifetime detection. We present a fiber-based method to obtain spectral information from a multiphoton fluorescence lifetime imaging (FLIM) system. This is made possible using the time delay introduced in the fluorescence emission path by a dispersive optical fiber coupled to a detector operating in time-correlated single-photon counting mode. This add-on spectral implementation requires only a few simple modifications to any existing FLIM system and is considerably more cost-efficient compared to currently available spectral detectors.

Cortex-wide neural interfacing via transparent polymer skulls.

Neural computations occurring simultaneously in multiple cerebral cortical regions are critical for mediating behaviors. Progress has been made in understanding how neural activity in specific cortical regions contributes to behavior. However, there is a lack of tools that allow simultaneous monitoring and perturbing neural activity from multiple cortical regions. We engineered 'See-Shells'-digitally designed, morphologically realistic, transparent polymer skulls that allow long-term (>300 days) optical access to 45 mm2 of the dorsal cerebral cortex in the mouse. We demonstrate the ability to perform mesoscopic imaging, as well as cellular and subcellular resolution two-photon imaging of neural structures up to 600 µm deep. See-Shells allow calcium imaging from multiple, non-contiguous regions across the cortex. Perforated See-Shells enable introducing penetrating neural probes to perturb or record neural activity simultaneously with whole cortex imaging. See-Shells are constructed using common desktop fabrication tools, providing a powerful tool for investigating brain structure and function.