Preprint FZJ-2026-04629

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png
Samples, Sources, Space: Decomposing Data Scale in Spatially Structured Representation Learning of Human Brain Microarchitecture

 ;  ;  ;  ;

2026
arXiv

arxiv () [10.48550/arXiv.2609.31201]

This record in other databases:  

Please use a persistent id in citations: doi:  doi:

Report No.: arXiv:2609.31201

Abstract: Scaling studies typically represent training data by a single count of samples. For hierarchically and spatially structured data, however, the same number of samples can be drawn from few or many sources and distributed differently across the underlying domain. We therefore study data scaling as an allocation problem, separating unique sample count, source diversity, and spatial coverage. We study this decomposition in microscopic whole-brain histology, where a source is an individual brain, and a sample is an image patch at a specific spatial location. Across 93 controlled pretraining runs of a contrastive model that uses spatial proximity for supervision, we vary data allocation, compute, and model capacity over 11.6 million spatially anchored image patches from 21 human brains. Performance improves with more unique samples, broader spatial coverage, additional compute, and larger model capacity. At fixed sample count, distributing samples across one to 18 subjects produces no detectable improvement, even though representations generalize substantially better to subjects encountered during pretraining. Inter-subject variation therefore strongly affects generalization, but additional subjects provide no benefit when a fixed sample budget is distributed across more sources. These results establish sample count, source diversity, and spatial coverage as distinct axes of data scaling in spatially structured representation learning.

Keyword(s): Artificial Intelligence (cs.AI) ; FOS: Computer and information sciences


Contributing Institute(s):
  1. Strukturelle und funktionelle Organisation des Gehirns (INM-1)
Research Program(s):
  1. 5254 - Neuroscientific Data Analytics and AI (POF4-525) (POF4-525)
  2. 5251 - Multilevel Brain Organization and Variability (POF4-525) (POF4-525)
  3. EBRAINS 2.0 - EBRAINS 2.0: A Research Infrastructure to Advance Neuroscience and Brain Health (101147319) (101147319)
  4. HIBALL - Helmholtz International BigBrain Analytics and Learning Laboratory (HIBALL) (InterLabs-0015) (InterLabs-0015)
  5. X-BRAIN (ZT-I-PF-4-061) (ZT-I-PF-4-061)
  6. DFG project G:(GEPRIS)501864659 - NFDI4BIOIMAGE - Nationale Forschungsdateninfrastruktur für Mikroskopie und Bildanalyse (NFDI4BIOIMAGE) (501864659) (501864659)

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

The record appears in these collections:
Institutssammlungen > INM > INM-1
Dokumenttypen > Berichte > Vorabdrucke
Workflowsammlungen > Öffentliche Einträge
Publikationsdatenbank
Open Access

 Datensatz erzeugt am 2026-09-28, letzte Änderung am 2026-11-05


OpenAccess:
Volltext herunterladen PDF
Externe links:
Volltext herunterladenVolltext
Volltext herunterladenFulltext by arXiv.org
Dieses Dokument bewerten:

Rate this document:
1
2
3
 
(Bisher nicht rezensiert)