Bringas Vega, M. L. L. | Search bioRxiv

bioRxiv2026

Source-space precision charts for lifespan EEG connectomics

Source-space electroencephalography (EEG) connectomics aims to estimate interactions among cortical generators from sensor cross-spectra that are mixed by the head and lead field. This task is difficult because marginal source covariance or coherence can retain leakage, common drive and indirect mediation, whereas developmental mapping requires conditional interactions that can be estimated repeatedly across large cohorts. We developed JSPACE (Joint Source-space Precision And Cross-spectral Estimation), a frequency-domain inverse framework for estimating multi-frequency source precision matrices from scalp cross-spectra. JSPACE couples posterior source cross-spectral estimation with standardized precision fitting, sparse frequency-smooth anatomical regularization, stochastic active-set optimization and post-selection refitting. In simulations, its advantage was target-specific: JSPACE reduced coherence inflation and achieved the lowest imaginary-coherence and peak-frequency errors in a forward neural-mass benchmark. When the ground-truth precision matrix was known, it achieved the highest exact, edge-collapsed and leakage-aware support recovery. We applied JSPACE to HarMNqEEG cross-spectral data from 1,935 participants aged 5.17-97.00 years, spanning 47 frequency bins and 360 cortical parcels. Affine-invariant Karcher tangent harmonization reconstructed subject-level estimates into a lifespan atlas of 360 diagonal and 64,620 source-pair age-frequency surfaces. The atlas revealed a continuous off-diagonal morphology landscape, in which age direction, frequency preference and interaction strength varied as overlapping axes rather than discrete edge classes. In contrast, diagonal precision surfaces shared a conserved alpha-trough morphology across parcels. Representative real-precision pathways captured posterior parietal, sensorimotor-parietal, frontopolar and visual-parietal motifs. Delta-band gradients were moderately aligned with the sensorimotor-association (S-A) organization of cortex from childhood through late adulthood, with a candidate oldest-old deviation in the sparsest age range. JSPACE provides a scalable framework for frequency-resolved source-precision charting in lifespan EEG.

neuroscience↗

bioRxiv2026

Generalized Normative Modeling: A One-Step Hierarchical Kernel Framework for Multi-Site Brain Charts with Self-Correcting Z-Scores

Normative modeling expresses individual brain phenotypes as z-scores relative to a population norm, but in multi-site studies batch effects contaminate these z-scores and undermine their biomarker value. Existing approaches either harmonize data before fitting a normative model (ComBat+Normative), letting residual site effects leak into z-scores, or use parametric one-step methods (GAMLSS, HBR) that cannot flexibly model multivariate covariate interactions. We propose Generalized Normative Modeling (GNM), a onestep hierarchical framework that jointly estimates the global trajectory and site-specific effects via NUFFT-accelerated kernel regression with GCV bandwidth selection. Because the z-score is the ratio of batch-corrected residual to batch-corrected scale, residual site variance cancels algebraically -- a property we term self-correction. On ABIDE I cortical thickness (387 HC, 11 sites, 68 ROIs) and HarMNqEEG log-power spectra (1,564 subjects, 14 sites, 18 channels x 235 frequency bins), GNM produced the most site-invariant z-scores and best age-signal preservation among four methods. This work provides an open-source MATLAB toolbox with a declarative formula interface (https://github.com/LMNonlinear/Generalized-Normative-Modeling), enabling reliable individual-level inference in pooled multi-site cohorts and advancing the use of normative deviations as clinical biomarkers in precision psychiatry and neurology.

bioengineering↗