Spectral feature modeling with graph signal processing for brain connectivity in autism spectrum disorder

• Graph signal processing models spectral brain connectivity in autism spectrum disorder (ASD).¹
• Methods are transferable to neural time series and connectivity analysis for decoding and biomarker work.¹
• Published in Nature; methods-focused with psychiatric context.^{1 1}

Weekly enrichment (2026-07-20)

• The referenced paper (Scientific Reports; Jabbar et al., Shenzhen University; published 2 July 2025) proposed a graph signal processing (GSP) framework combining spectral-domain and topological features to model brain connectivity in autism spectrum disorder.^{2 3}
• The method built subject-specific connectivity graphs (nodes = brain regions, edges = functional interactions) from publicly available fMRI and EEG datasets sourced from Kaggle.^{2 3}
• Extracted features included Graph Fourier Transform coefficients, spectral entropy, and clustering coefficients, fused via Principal Component Analysis and classified with a support vector machine using an RBF kernel.^{2 3}
• The paper reported 98.8% classification accuracy, and a feature-ablation analysis found that removing spectral entropy dropped performance by roughly 30%.²
• A 25% sparsity threshold in graph construction was reported to balance robustness and computational efficiency.²
• Critical caveat: this article was RETRACTED on 11 May 2026 for missing information about the features used and inconsistent sample-size reporting throughout.⁴
• Because subject counts were inconsistently reported, the exact sample sizes are not reliably reported and the headline accuracy should be treated as unverified.⁴
• Methodologically, GSP remains an established approach in adjacent, non-retracted work that integrates structural and functional brain graphs for ASD classification.³

Footnotes

1. https://news.google.com/rss/articles/CBMiX0FVX3lxTE9OcWFpQ3Awd0ZFVDF1U0hqeEYwQ1BEcTAzdm1UMmI2a09CV0lWMmVEdEFqYkFJb05yb0lLYTJzZ2xKaTdkUF9lbWdSdVB6b2JHeC1yQi0yY2RtZXdueTJr?oc=5 ↩ ↩² ↩³ ↩⁴

2. https://www.nature.com/articles/s41598-025-06489-6 ↩ ↩² ↩³ ↩⁴ ↩⁵

3. https://pmc.ncbi.nlm.nih.gov/articles/PMC12217149/ ↩ ↩² ↩³ ↩⁴

4. https://doi.org/10.1038/s41598-026-51646-0 ↩ ↩²