Vitreoretinal Surgical Instrument Tracking in Three Dimensions Using Deep Learning.

Purpose: To evaluate the potential for artificial intelligence-based video analysis to determine surgical instrument characteristics when moving in the three-dimensional vitreous space. Methods: We designed and manufactured a model eye in which we recorded choreographed videos of many surgical instruments moving throughout the eye. We labeled each frame of the videos to describe the surgical tool characteristics: tool type, location, depth, and insertional laterality. We trained two different deep learning models to predict each of the tool characteristics and evaluated model performances on a subset of images. Results: The accuracy of the classification model on the training set is 84% for the x-y region, 97% for depth, 100% for instrument type, and 100% for laterality of insertion. The accuracy of the classification model on the validation dataset is 83% for the x-y region, 96% for depth, 100% for instrument type, and 100% for laterality of insertion. The close-up detection model performs at 67 frames per second, with precision for most instruments higher than 75%, achieving a mean average precision of 79.3%. Conclusions: We demonstrated that trained models can track surgical instrument movement in three-dimensional space and determine instrument depth, tip location, instrument insertional laterality, and instrument type. Model performance is nearly instantaneous and justifies further investigation into application to real-world surgical videos. Translational Relevance: Deep learning offers the potential for software-based safety feedback mechanisms during surgery or the ability to extract metrics of surgical technique that can direct research to optimize surgical outcomes.

Spatiotemporal Mapping of Brain Cilia Reveals Region-Specific Oscillation of Length and Orientation.

In this study, we conducted high-throughput spatiotemporal analysis of primary cilia length and orientation across 22 mouse brain regions. We developed automated image analysis algorithms, which enabled us to examine over 10 million individual cilia, generating the largest spatiotemporal atlas of cilia. We found that cilia length and orientation display substantial variations across different brain regions and exhibit fluctuations over a 24-hour period, with region-specific peaks during light-dark phases. Our analysis revealed unique orientation patterns of cilia at 45 degree intervals, suggesting that cilia orientation within the brain is not random but follows specific patterns. Using BioCycle, we identified circadian rhythms of cilia length in five brain regions: nucleus accumbens core, somatosensory cortex, and three hypothalamic nuclei. Our findings present novel insights into the complex relationship between cilia dynamics, circadian rhythms, and brain function, highlighting cilia crucial role in the brain's response to environmental changes and regulation of time-dependent physiological processes.