siibra - Software interface for interacting with brain atlases

Copyright 2020-2022, Forschungszentrum Jülich GmbH

Authors: Big Data Analytics Group, Institute of Neuroscience and Medicine (INM-1), Forschungszentrum Jülich GmbH

siibra is a Python client for working with brain atlas frameworks that integrate multiple brain parcellations and reference spaces across different spatial scales, and connect them with a multimodal regional data features. It aims to facilitate the programmatic and reproducible incorporation of brain region features from different sources into reproducible neuroscience workflows.

This repository implements a basic commandline interface (CLI) for siibra.

Note: ``siibra-cli`` is still in development. While care is taken that it works reliably, its API is not yet stable and you may still encounter bugs when using it.

siibra provides structured acccess to parcellation schemes in different brain reference spaces, including volumetric reference templates at both macroscopic and microscopic resolutions as well as surface representations. It supports both discretely labelled and continuous (probabilistic) parcellation maps, which can be used to assign brain regions to spatial locations and image signals, to retrieve region-specific neuroscience datasets from multiple online repositories, and to sample information from high-resolution image data. Among the datasets anchored to brain regions are many different modalities from in-vivo and post mortem studies, including regional information about cell and transmitter receptor densties, structural and functional connectivity, gene expressions, and more.

siibra is mainly developed by the Human Brain Project for accessing the EBRAINS human brain atlas. It stores much of its contents in the EBRAINS Knowledge Graph, and is designed to support the OpenMINDS metadata standards. Its functionalities include common actions known from the interactive viewer siibra explorer hosted on EBRAINS. In fact, the viewer is a good resource for exploring siibra’s core functionalities interactively: Selecting different parcellations, browsing and searching brain region hierarchies, downloading maps, identifying brain regions, and accessing multimodal features and connectivity information associated with brain regions. Feature queries in siibra are parameterized by data modality and anatomical location, while the latter could be a brain region, brain parcellation, or location in reference space. Beyond the functionality of siibra-explorer, the Python library also supports a range of data analysis features suitable for typical neuroscience workflows.

siibra hides much of the complexity that would be required to collect and interact with the individual paracellations,templates andd data repositories. By encapsulating many aspects of interacting with different maps and reference templates spaces, it also minimizes common errors like misinterpretation of coordinates from different reference spaces, mixing up label indices of brain regions, or utilisation of inconsistent versions of parcellation maps. It aims to provide a safe way of using maps defined across multiple spatial scales for reproducible analysis.