Event Abstract

Back to Event

Mojo 2.0: Connectome Annotation Tool

Seymour Knowles-Barley1*, Mike Roberts2, Narayanan Kasthuri1, Dongil Lee1, Hanspeter Pfister2 and Jeff W. Lichtman1
  • 1 Harvard University, Department of Molecular and Cellular Biology, United States
  • 2 Harvard University, School of Engineering and Applied Sciences, United States

A connectome is the wiring diagram of connections in a nervous system. Mapping this network of connections is necessary for discovering the underlying architecture of the brain and investigating the physical underpinning of cognition, intelligence, and consciousness [1, 2, 3]. It is also an important step in understanding how connectivity patterns are altered by mental illnesses, learning disorders, and age related changes in the brain.

Mapping the densely packed network of synaptic connections between individual neurons in the brain is challenging, even when using state-of-the-art microscopy techniques, computer vision algorithms, and analysis software. Using serial section electron microscopy (SSEM), it is possible to image of thousands of neuronal profiles and synaptic connections at the nm scale and acquire terabytes (TB) of image data. Fully automatic computer vision techniques are available to annotate this data, but results are still far from perfect and require additional human annotation to produce an accurate connectivity map [4]. Therefore, it is important to develop tools that allow domain experts to interact with the raw data and ensure correctness. A number of software tools and packages for neuron reconstruction are available [5, 6, 7, 8, 9, 10]. However, tools supporting dense reconstruction and correction of automatic segmentations over TB datasets are under development. User interaction modes are an important consideration for these tools to reduce the amount manual labour required to produce a 3D reconstruction.

In this demonstration we present Mojo 2.0, an open source, interactive, scalable annotation tool to correct errors in automatic segmentation results (Figure 1). Sparse user scribbles are used to correct both split and merge errors for TB scale volumes.

To correct a split error, the user scribbles over objects that should be joined. Any segments touched by the brush stroke are combined into one segment. This operation is performed in 2D or 3D as specified by the user. To correct a merge error the user can use a split brush to paint a broad line over cell membrane in the image. Watershed pixels within this region are used to split the segment in 2D. Mojo 2.0 also predicts how adjacent 2D segments should be split, allowing the user to navigate in 3D and quickly confirm or adjust multiple split operations while retaining 3D connectivity. An adjust mode allows the user to manually draw a region and add it to the selected segment. This mode is useful when a combination of split and merge operations are required to correct a segment.

We demonstrate Mojo 2.0 on SSEM images of mouse brain cortex, automatically labeled by the Rhoana image processing pipeline and compare results with manual tracing methods.

Mojo 2.0 is available online at: http://www.rhoana.org/



Figure 1: Mojo 2.0 is an interactive desktop application for connectome annotation. The Mojo 2.0 interface shown here displays electron microscope images of mouse cortex with segmentations shown as an editable color overlay. Automatic segmentation results can be corrected by sparse user scribbles. Above the editing panel a simple toolbar controls display and annotation options. On the right a list of segments is displayed, ordered by segment size. The application is open source and scalable up to TB scale volumes.

Figure 1
Image

References

[1] Narayanan Kasthuri and Jeff W. Lichtman. Neurocartography. Neuropsychopharmacology, 35(1):342–343, 2010.

[2] Jeff W. Lichtman and Joshua R. Sanes. Ome sweet ome: what can the genome tell us about the connectome? Current Opinion in Neurobiology, 18(3):346–353, 2008.

[3] H Sebastian Seung. Reading the book of memory: Sparse sampling versus dense mapping of connectomes. Neuron, 62(1):17–29, 2009.

[4] Verema Kaynig, Amelio Vazquez-Reina, Seymour Knowles-Barley, Mike Roberts, Thouis R. Jones, Narayanan Kasthuri, Eric Miller, Jeff W. Lichtman, and Hanspeter Pfister. Large Scale Automatic Reconstruction of Neuronal Processes from Electron Microscopy Images. IEEE Transactions on Medical Imaging (under review).

[5] Albert Cardona. TrakEM2: an ImageJ-based program for mophological data mining and 3D modeling. In Proceedings of the ImageJ User and Developer Conference, pages 43–51, 2006.

[6] J. C. Fiala. Reconstruct: a free editor for serial section microscopy. Journal of microscopy, 218(Pt 1):52–61, April 2005.

[7] Moritz Helmstaedter, Kevin L Briggman, and Winfried Denk. High-accuracy neurite reconstruction for high-throughput neuroanatomy. Nature Neuroscience, 14(8):1081–1088, 2011.

[8] Won-Ki Jeong, Johanna Beyer, Markus Hadwiger, Rusty Blue, Charles Law, Amelio V´azquez-Reina, Clay Reid, Jeff Lichtman, and Hanspeter Pfister. SSECRETT and NeuroTrace: Interactive visualization and analysis tools for large-scale neuroscience datasets. IEEE Computer Graphics and Applications, 30:58–70, 2010.

[9] Mike Roberts, Won-Ki Jeong, Amelio V´azquez-Reina, Markus Unger, Horst Bischof, Jeff Lichtman, and Hanspeter Pfister. Neural process reconstruction from sparse user scribbles. In Medical Image Computing and Computer Assisted Intervention (MICCAI ’11), pages 621–62, 2011.

[10] Christoph Sommer, Christoph N. Straehle, Ullrich K¨othe, and Fred A. Hamprecht. Ilastik: Interactive learning and segmentation toolkit. In Proceedings of the 8th International Symposium on Biomedical Imaging (ISBI 2011), pages 230–233.

Keywords: Mojo, connectome, connectomics, connectome mapping, annotation, Electron microscopy, Computational Neuroanatomy, human computer interaction, big data

Conference: Neuroinformatics 2013, Stockholm, Sweden, 27 Aug - 29 Aug, 2013.

Presentation Type: Demo

Topic: Neuroimaging

Citation: Knowles-Barley S, Roberts M, Kasthuri N, Lee D, Pfister H and Lichtman JW (2013). Mojo 2.0: Connectome Annotation Tool. Front. Neuroinform. Conference Abstract: Neuroinformatics 2013. doi: 10.3389/conf.fninf.2013.09.00060

Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.

The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.

Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.

For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.

Received: 29 Apr 2013; Published Online: 11 Jul 2013.

* Correspondence: Dr. Seymour Knowles-Barley, Harvard University, Department of Molecular and Cellular Biology, Cambridge, MA, United States, seymourkb@seas.harvard.edu