Table 1_SliceMap: A Binary Classification Pipeline for Brain MRI Region Detection

A 5.7 MB document by Xiaeya Jiang, last updated May 15, 2026, describes a performance-guided 2D slice-based pipeline for identifying candidate spatial regions in brain MRI. The pipeline uses binary classification with convolutional neural networks trained on plane-specific slices and occlusion-based attribution analysis. It was demonstrated using sex classification to identify a candidate region consistent with the anterior cingulate cortex and adjacent medial structures.

Use Cases

• Developing binary classification models for brain MRI based on anatomical plane slices.
• Identifying spatially localized neuroanatomical signals for disease mechanism investigation.
• Evaluating slice-level performance for candidate region localization in high-dimensional MRI data.
• Applying occlusion-based attribution analysis to interpret model decisions in medical imaging.

Strengths

• The pipeline is demonstrated on a specific, controlled testbed (sex classification) with results consistent with prior neuroanatomical findings.
• The document is shared under a permissive CC-BY-4.0 license.
• The method is designed to address the challenge of detecting subtle, spatially localized signals from high-dimensional MRI data.

Limitations

• The dataset's row count and specific column structure are unknown, limiting suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The data is a 5.7 MB DOCX file, which is a small dataset and may contain primarily methodological description rather than raw data.

Provenance

Source

figshare

Collection Method

Proposed as a methodological pipeline for brain MRI analysis.

Freshness

Last updated 2026-05-15 04:26:26; freshness should be verified.