The impact of extended reality simulators on ophthalmic surgical training and performance: a systematic review and meta-analysis of 17,623 eyes
Related Articles
Immersive auditory-cognitive training improves speech-in-noise perception in older adults with varying hearing and working memory
Ageing is associated with elevated pure-tone thresholds, accompanied by increased difficulties in understanding speech-in-noise. While amplification provides important, but insufficient support, auditory-cognitive training (ACT) might propose a solution. However, generalized effects have been scarce, highlighting the necessity of training designs targeting naturalistic listening situations. We addressed this issue by designing a short-term ACT in a purely auditory- and a virtual multisensory environment, targeting both, sensory and cognitive processing of natural speech. 40 healthy older participants with varying hearing- and cognitive capacities were exposed to both trainings (cross-over design), while speech-in-noise perception was measured before and after each session. Immersive ACT exposure resulted in increased speech-in-noise perception, particularly for individuals with more pronounced hearing loss or reduced auditory working memory capacity. These results demonstrate that combining sensory and cognitive training elements, particularly in a multisensory environment, has the potential to improve speech in noise perception.
Surgical video workflow analysis via visual-language learning
Surgical video workflow analysis has made intensive development in computer-assisted surgery by combining deep learning models, aiming to enhance surgical scene analysis and decision-making. However, previous research has primarily focused on coarse-grained analysis of surgical videos, e.g., phase recognition, instrument recognition, and triplet recognition that only considers relationships within surgical triplets. In order to provide a more comprehensive fine-grained analysis of surgical videos, this work focuses on accurately identifying triplets <instrument, verb, target> from surgical videos. Specifically, we propose a vision-language deep learning framework that incorporates intra- and inter- triplet modeling, termed I2TM, to explore the relationships among triplets and leverage the model understanding of the entire surgical process, thereby enhancing the accuracy and robustness of recognition. Besides, we also develop a new surgical triplet semantic enhancer (TSE) to establish semantic relationships, both intra- and inter-triplets, across visual and textual modalities. Extensive experimental results on surgical video benchmark datasets demonstrate that our approach can capture finer semantics, achieve effective surgical video understanding and analysis, with potential for widespread medical applications.
Evaluation of electrical impedance spectroscopy of bovine eyes for early detection of uveal melanoma
Uveal melanoma is the most common primary intraocular cancer in adults and is an aggressive malignancy with risk to vision and survival. Early detection and timely management of tumors may help preserve vision and reduce mortality rate but is challenging as many tumors are asymptomatic until they become large. Here, we studied the electrical properties of eyes to investigate a novel method for potentially detecting small intraocular tumors. We used finite element analysis to simulate the impact of uveal melanoma tumors on electrical impedance and current density in eye models. We also measured the impedance and current flow in the presence of inserted tissue simulating an intraocular tumor in enucleated bovine eyes and eyes in bovine head ex vivo. Our results showed that a 5 mm-diameter mass was detected inside a 32-mm diameter bovine eye by the impedance analyzer.
Faster implicit motor sequence learning of new sequences compatible in terms of movement transitions
New information that is compatible with pre-existing knowledge can be learned faster. Such schema memory effect has been reported in declarative memory and in explicit motor sequence learning (MSL). Here, we investigated if sequences of key presses that were compatible to previously trained ones, could be learned faster in an implicit MSL task. Participants trained a motor sequence before switching to a completely new sequence, to a compatible sequence with high overlap in ordinal positions, or to an incompatible sequence with low overlap, while the compatible and incompatible sequences had the same overlap in movement transitions. We observed accelerated learning in the Compatible and Incompatible groups compared to the New group, if participants trained for 3 sessions before switching to the altered sequence. Overall, our study suggests facilitative learning of implicit motor sequences that are compatible in movement transitions, if the previous sequence has been trained extensively.
Cellpose3: one-click image restoration for improved cellular segmentation
Generalist methods for cellular segmentation have good out-of-the-box performance on a variety of image types; however, existing methods struggle for images that are degraded by noise, blurring or undersampling, all of which are common in microscopy. We focused the development of Cellpose3 on addressing these cases and here we demonstrate substantial out-of-the-box gains in segmentation and image quality for noisy, blurry and undersampled images. Unlike previous approaches that train models to restore pixel values, we trained Cellpose3 to output images that are well segmented by a generalist segmentation model, while maintaining perceptual similarity to the target images. Furthermore, we trained the restoration models on a large, varied collection of datasets, thus ensuring good generalization to user images. We provide these tools as ‘one-click’ buttons inside the graphical interface of Cellpose as well as in the Cellpose API.
Responses