Embedded Files
Aim 1: Microendoscope-specific signal enhancement and motion artifact correction
Aim 2: Machine-learning implementation to enable real-time feature segmentation
Aim 3: Correlative live-and-postmortem imaging for cell activity and molecular co-profiling
Microendoscopic Optimization
Microendoscope imaging in live animals suffers from low signal-to-noise ratio (SNR), motion artifacts from respiration, heart beating and organ peristalsis, small field of view (FOV), and limited volume.We will improve SNR by training an AI model with paired 1 fps and 50 fps images acquired with the same FOV from tissues ex vivo. The ex vivo, motion artifact-free, 1 fps images have very high SNR, which can be used as a training target for 50 fps images to enhance image quality. The faster frame rate will greatly benefit motion artifact correction in vivo. For motion correction, blur frames will be automatically detected and potentially replaced by AI-generated images. Panorama and 3D stitching will be used to extend the FOV and volume.Microendoscopic Optimization
AI Real-Time Segmentation
Generating segmentation inputs for machine-learning model training has been a laborious and error-prone process for biomedical researchers. On the other hand, current microendoscope post-acquisition processing/analysis pipelines work on images after they are stored in HDD/SSD. These prevent real-time image artifact correction and annotation for better user experience or prompt decision-making in clinical settings. We show that a human-supervised strategy can permit rapid AI model training in 3-5 rounds of iterations with minimal human efforts. We recently also developed software (SNDiF) to handle a 1.1 Gigavoxel/sec imaging data flow. We will adapt the same image buffering and processing principle to integrate GPU-implemented machine-learning models for real-time cell and tissue feature segmentation.AI Real-Time Segmentation
Correlation Imaging
Current microendoscopes can image a limited number of labeled probes with limited optical resolution and throughput. High-plex molecular information that correlates to live activity imaging with single-cell resolution is lacking. Imaging and wet-lab protocols will be developed to monitor individual cellular activity in live animals over time, followed by thick tissue clearing and multi-plex mRNA/protein co-profiling in ex vivo tissues. We will also optimize computation pipelines to register the same cells in the live and postmortem images for activity and molecular co-profiling.Correlation Imaging
Tissue microenvironment. Microendoscopes will be equipped with monolithic, 3-axis scanning mechanisms for use to collect fluorescence images in vivo from the epithelium in A) horizontal (XY) and B) vertical (XZ) planes. C) These images will be combined to generate 3D volumetric image. Image processing methods, including handcrafted and deep learning, will be used to visualize cell-cell interactions in the tissue microenvironment.
Publications
Letner JG, Patel PR, Hsieh JC, Smith Flores IM, Della Valle E, Walker LA, Weiland JD, Chestek CA, Cai D. Post-explant profiling of subcellular-scale carbon fiber intracortical electrodes and surrounding neurons enables modeling of recorded electrophysiology. J Neural Eng 2023;20:026019. PMC10022369.
Farzaneh N, Stein EB, Soroushmehr R, Gryak J, Najarian K. A deep learning framework for automated detection and quantitative assessment of liver trauma. BMC Med Imaging 2022;22:39. PMC8905785.
Minoccheri C, Williamson CA, Hemmila M, Ward K, Stein EB, Gryak J, Najarian K. An interpretable neural network for outcome prediction in traumatic brain injury. BMC Med Inform Decis Mak 2022;22:203. PMC9341077.
Walker LA, Williams JS, Li Y, Roossien DH, Lee WJ, Michki NS, Cai D. nGauge: Integrated and Extensible Neuron Morphology Analysis in Python. Neuroinformatics 2022; PMC9720862.
Kim RB, Alge OP, Liu G, Biesterveld BE, Wakam G, Williams AM, Mathis MR, Najarian K, Gryak J. Prediction of postoperative cardiac events in multiple surgical cohorts using a multimodal and integrative decision support system. Sci Rep 2022;12:11347. PMC9256604.
Yao H, Najarian K, Gryak J, Bishu S, Rice MD, Waljee AK, Wilkins HJ, Stidham RW. Fully automated endoscopic disease activity assessment in ulcerative colitis. Gastrointest Endosc 2021;93:728-736.e1. PMID: 32810479.
Li Y, Walker LA, Zhao Y, Edwards EM, Michki NS, Cheng HPJ, Ghazzi M, Chen TY, Chen M, Roossien DH, Cai D. Bitbow Enables Highly Efficient Neuronal Lineage Tracing and Morphology Reconstruction in Single Drosophila Brains. Front Neural Circuits 2021;15:732183. PMC8564373.
Iyer K, Najarian CP, Fattah AA, Arthurs CJ, Soroushmehr SMR, Subban V, Sankardas MA, Nadakuditi RR, Nallamothu BK, Figueroa CA. AngioNet: a convolutional neural network for vessel segmentation in X-ray angiography. Sci Rep 2021;11:18066. PMC8433338.
Kim RB, Gryak J, Mishra A, Cui C, Soroushmehr SMR, Najarian K, Wrobel JS. Utilization of smartphone and tablet camera photographs to predict healing of diabetes-related foot ulcers. Comput Biol Med 2020;126:104042. PMC9058995.
Shen FY, Harrington MM, Walker LA, Cheng HPJ, Boyden ES, Cai D. Light microscopy based approach for mapping connectivity with molecular specificity. Nature Communications 2020;11: 4632. PMC7493953.
Patel PR, Popov P, Caldwell CM, Welle E, Egert D, Pettibone JR, Roossien, DH, Becker, J, Berke, J, Chestek, CA#, Cai, D. High density carbon fiber arrays for chronic electrophysiology, fast scan cyclic voltammetry, and correlative anatomy. Journal of Neural Engineering 2020;17:056029. PMID: 33055366.
Egert D, Pettibone JR, Lemke S, Patel PR, Caldwell CM, Cai D, Ganguly K, Chestek CA, Berke J (2020) Cellular-scale silicon probes for high-density, precisely-localized neurophysiology. Journal of Neurophysiology 124:1578. PMC7814906.
Reamaroon N, Sjoding MW, Derksen H, Sabeti E, Gryak J, Barbaro RP, Athey BD, Najarian K. Robust segmentation of lung in chest x-ray: applications in analysis of acute respiratory distress syndrome. BMC Med Imaging 2020;20:116. PMC7566051.
Bianchi J, de Oliveira Ruellas AC, Gonçalves JR, Paniagua B, Prieto JC, Styner M, Li T, Zhu H, Sugai J, Giannobile W, Benavides E, Soki F, Yatabe M, Ashman L, Walker D, Soroushmehr R, Najarian K, Cevidanes LHS. Osteoarthritis of the Temporomandibular Joint can be diagnosed earlier using biomarkers and machine learning. Sci Rep 2020;10:8012. PMC7228972.
Shen FY, Harrington MM, Walker LA, Cheng HPJ, Boyden ES, Cai D (2020) Light microscopy based approach for mapping connectivity with molecular specificity. Nature Communications 11:4632. PMC7493953.
Yao H, Williamson C, Gryak J, Najarian K. Automated hematoma segmentation and outcome prediction for patients with traumatic brain injury. Artif Intell Med 2020;107:101910. PMID: 32828449.
Dizaji AS, Walker LA, Cai D (2020) TraceMontage: A method for merging multiple independent neuronal traces. J Neurosci Methods 332:108560. PMC6950642.
Roossien DH, Sadis BV, Yan Y, Webb JM, Min LY, Dizaji AS, Bogart LJ, Mazuski C, Huth RS, Stecher JS, Akula S, Shen F, Li Y, Xiao T, Vandenbrink M, Lichtman JW, Hensch TK, Herzog ED, Cai D. Multispectral tracing in densely labeled mouse brain with nTracer. Bioinformatics 2019;35:3544-46. PMC6748755.
Yao H, Stidham RW, Soroushmehr R, Gryak J, Najarian K. Automated Detection of Non-Informative Frames for Colonoscopy Through a Combination of Deep Learning and Feature Extraction. Annu Int Conf IEEE Eng Med Biol Soc 2019;2019:2402-2406. PMID: 31946383.
Veling MW, Li Y, Veling MT, Litts C, Michki N, Liu H, Ye B, Cai D. Identification of Neuronal Lineages in the Drosophila Peripheral Nervous System with a "Digital" Multi-spectral Lineage Tracing System. Cell Rep. 2019;29:3303-3312.e3. PMC6913890.
Report abuse