Poster (After Call) FZJ-2026-06030

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png
Integrated Molecular, Structural, and Functional Brain Atlases in the NiSpace Toolbox

 ;  ;  ;

2026

Organization for Human Brain Mapping (OHBM), BordeauxBordeaux, France, 14 Jun 2026 - 18 Jun 20262026-06-142026-06-18 () [10.34734/FZJ-2026-06030]

This record in other databases:  

Please use a persistent id in citations: doi:

Abstract: IntroductionMultimodal brain maps derived from nuclear imaging, structural and functional MRI, spatial transcriptomics, and histology are increasingly used as reference for spatial inference methods, aiming to better understand MRI-based neuroimaging findings1. Yet these resources are typically distributed in heterogeneous formats, imaging spaces, and parcellations, which complicates analysis and reproducibility. NiSpace (NeuroImaging Spatial Colocalization Environment; https://github.com/leondlotter/nispace) is a Python-based toolbox to assess spatial correlations ("colocalizations") between brain maps2. It aims to be the most comprehensive, yet easy-to-use and flexible framework for reference data retrieval, colocalization estimation, and significance testing to date. NiSpace extends the widely used neuromaps toolbox3 in that it provides high-level workflow functionalities and advanced statistical methods, condensing, for example, a spatial correlation analysis with 48 different PET maps or an imaging-based gene-set enrichment analysis using the Allen Human Brain Atlas (AHBA) in a single command each. Here, we introduce the curated library of reference datasets currently included with NiSpace. We describe how the data are collected, standardized, and distributed, together with an exemplary validation of the standardization procedure.MethodsWe aim to include any publicly available dataset of group-level brain maps that is of interest to a broad range of users. To be included, a dataset has to have a license permitting modification and redistribution, provide a whole‑brain or whole-cortex coverage at the group level, and must be available in at least one of the currently supported standard imaging spaces (MNI152, fsaverage, fsLR) or in at least one of the currently supported brain parcellations (see Fig. 1). From the source dataset, we generate NiSpace reference datasets in as many of the currently supported imaging spaces as possible (MNI152NLin6Asym, MNI152NLin2009cAsym, fsaverage, fsLR). All data are stored in a GitHub repository separate to the library's main repository, referenced through stable identifiers, and accessed via the NiSpace Python API as either imaging files or parcellated table data. To ease straight-forward retrieval and analysis of data according to user preferences, and to facilitate set enrichment analyses, we additionally provide "collections" of maps within a given dataset. A collection subsets and/or groups maps into reasonable categories. This design aims to minimize manual handling of file formats, imaging spaces, and lookup tables. The nuclear imaging map collection ("pet") represents a special case in that the included maps are both the most used4,5 but also the most heterogenous. Extending the nuclear imaging data provided with the neuromaps toolbox, NiSpace includes standardized versions of all maps, which were non-linearly co-registered with a deep learning assisted procedure (https://martinos.org/malte/synthmorph/), scaled, and masked. To demonstrate the advantage of our standardization procedure, we show how the voxel-wise correlation between three PET maps using the same tracer (ABP688, binding to the mGluR5 receptor) but obtained from independent samples improves after standardization.


Contributing Institute(s):
  1. Gehirn & Verhalten (INM-7)
Research Program(s):
  1. 5254 - Neuroscientific Data Analytics and AI (POF4-525) (POF4-525)

Appears in the scientific report 2026
Database coverage:
OpenAccess
Click to display QR Code for this record

The record appears in these collections:
Document types > Presentations > Poster
Institute Collections > INM > INM-7
Workflow collections > Public records
Publications database
Open Access

 Record created 2026-11-16, last modified 2026-11-16


OpenAccess:
Download fulltext PDF
Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)