Experimental Parameter Settings for fMRI-Based MDD Diagnosis Model

1,160 participants, including 597 patients with major depressive disorder and 563 healthy controls, provided resting-state fMRI data for a novel graph contrastive learning framework. The framework, developed by Dan Long and last updated in May 2026, achieved a diagnostic classification AUC of 71.2%. It identified key brain regions like the dorsolateral superior frontal gyrus, thalamus, and insula linked to MDD pathophysiology.

Use Cases

• Training graph neural network models for MDD diagnosis based on functional connectivity matrices.
• Benchmarking novel deep learning frameworks against conventional machine learning models for neuroimaging classification.
• Investigating the role of specific brain regions like the dorsolateral superior frontal gyrus in depression pathophysiology.
• Reproducing the experimental setup for the EC-GCL (Edge Convolution Graph Contrastive Learning) framework.

Strengths

• Dataset includes a substantial sample of 1,160 participants for a clinical neuroimaging study.
• Clear clinical labels are provided for 597 MDD patients and 563 healthy controls.
• The associated model achieved a reported AUC of 71.2%, providing a performance benchmark.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small at 5.5 KB, suggesting it contains only parameter settings, not the underlying fMRI data.

Provenance

Source

figshare

Collection Method

Likely contains experimental parameters for a model trained on rs-fMRI data from a clinical cohort.

Freshness

Last updated 2026-05-05 17:43:02; freshness should be verified.