Very excited to share that our DELiVR method is now open access published @NatureMethods. We created a simple, brain-wide cell analysis deep learning tool, no coding needed! Fiji Plugin makes it accessible to all. by Doris Kaltenecker Rami @moritz_negwer

2/n DELiVR is a robust deep-learning pipeline for whole-brain cell mapping. It operates through a user-friendly Fiji plugin. Re-trainable on custom data, DELiVR simplifies and streamlines brain cell analysis.

3/n Background: Tissue clearing and fluorescent imaging have transformed how we view protein activities in whole brains, offering a detailed map of neuronal and cellular dynamics across entire systems.

4/n Challenges persist: Imaging whole mouse brains produces large datasets that require reliable analysis. While AI has immense potential, its use is often limited to specialized labs due to the need for extensive annotation and advanced coding skills.

5/n DELiVR is our answer to these challenges. A deep learning solution for brain-wide cell mapping, using VR-based annotation. Plus, it's all wrapped in a simple Fiji plugin interface.

6/n The VR Advantage: A major bottleneck in training robust deep learning is the volume of required annotated data, which traditional 2D methods struggle to supply quickly. By adopting VR, we have accelerated & enhanced the reliability of the annotation process. #VirtualReality

7/n VR not only sped up annotation but also enhanced the quality of our training data, providing a robust dataset for training deep learning models.

8/n In addition, DELiVR outperforms traditional segmentation methods for c-Fos, a marker for neuronal activity! #MachineLearning

9/n DELiVR offers complete brain segmentation in both original image and atlas space. Cells are mapped and color-coded based on Area IDs for precise regional analysis. Check out our @syGlassVR for a detailed view! #Neuroscience

10/n Available for use on any machine from PCs to clusters, DELiVR's compatibility with Docker and Fiji ensures a seamless experience for all researchers. Just plug and play! 🔗

11/n After community feedback, we added a Docker container, that also runs via Fiji for researchers to tailor deep learning models to their specific needs. We highlight this feature by segmenting cell bodies of microglia, the immune cells of the brain.

12/n The result output of DELiVR allows users to easily observe the color-coded AI segmentation in the original image in different brain sub-areas. Perfect for validating the results! #Imaging

13/n Case Study: Using DELiVR, we explored brain activity differences in cancer-associated weight changes in mice, revealing unique insights into neurophysiological adaptations. #CancerResearch

14/n Findings Highlight: Mice with weight-stable cancer showed increased neuronal activity, unlike their weight-losing counterparts. This highlights a previously unknown neurophysiological phenotype in cancer-related weight control. #Cachexia

15/n Many hyperactive areas were pinpointed in the cortex, highlighting the complex brain-body interactions in cancer dynamics.

16/n In Summary: DELiVR is not just a tool but a revolution in analyzing neuronal activity with ease and precision, thanks to its VR-trained, deep learning pipeline and user-friendly Fiji plugin.

17/n Beyond neuroscience, DELiVR's application can be a template for other fields, showing how integrating VR with AI annotation can drastically improve data quality and analysis speed.

18/n Explore more about DELiVR, access tools, and see it in action! GitHub 🔗Videos & More #OpenScience

19/n Read the accompanying paper, out now in Nature Methods!

20/n This was a team effort! We couldn't have done it without @Lu_h99 @neuronflow @shan_heather @MihailMuc @zhouyi_rong @jocpae Benedikt Wiestler @MariePiraud @DanielRueckert @JuliaZuber08 @MorignyPauline @RohmMaria @menze_group Stephan Herzig @BerrielDiaz

21/n Work done at @HelmholtzMunich, @LMU_Uniklinikum, @ISD_Research, @TU_Muenchen, @LMU_Muenchen, @kocuniversity and supported by @ERC_Research, and amazing @syGlassVR software.