Altered individual-level morphological similarity brain networks in late-life subthreshold depression: Preserved global topology with selective nodal and sub-network disruptions

AI Summary

Global small world topology is preserved in late-life subthreshold depression despite network-level alterations.
Selective nodal disruptions include increased left hippocampal degree centrality and reduced efficiency in right inferior occipital and postcentral gyri, with a 50-node hypoconnectivity sub-network.
Occipitotemporal hypoconnectivity correlates with reminiscence and rumination; postcentral efficiency associates with cognition and extraversion, suggesting compensatory hippocampal reorganisation.

TL;DR (AI Generated)

Graph metrics were assessed across sparsity 0.05-0.40 with Benjamini-Hochberg false-discovery-rate (FDR) control, and the network-based statistic identified altered sub-networks (5000 permutations, FWE-corrected).

Participants with StD had increased left hippocampal degree centrality (p = 0.0013), reduced nodal efficiency in the right inferior occipital (p = 0.0046) and right postcentral gyri (p = 0.0085), and a 50-node, 73-edge hypoconnectivity sub-network (pFWE = 0.0408); none of these regional findings survived FDR correction.

CONCLUSIONS: Late-life StD shows preserved global topology with focal nodal and sub-network disruptions.

Basic summary

Int Psychogeriatr. 2026 Nov 13:100307. doi: 10.1016/j.inpsyc.2026.100307. Online ahead of print.

ABSTRACT

OBJECTIVES: Subthreshold depression (StD) is prevalent in older adults and increases the risk of major depressive disorder, but its large-scale brain-network organization remains poorly understood. Individual-level morphological similarity networks estimate interregional structural covariance separately for each participant.

METHODS: 110 older women (56 StD, 54 comparison participants) underwent T1-weighted MRI. Individual networks were constructed using Jensen-Shannon divergence-based similarity (JSDs) between gray matter volume distributions across 116 AAL regions. Graph metrics were assessed across sparsity 0.05-0.40 with Benjamini-Hochberg false-discovery-rate (FDR) control, and the network-based statistic identified altered sub-networks (5000 permutations, FWE-corrected). Reproducibility was tested with an alternative Kullback-Leibler metric (KLDs) and 4-mm smoothing.

RESULTS: Both groups showed preserved small-world global topology. Participants with StD had increased left hippocampal degree centrality (p = 0.0013), reduced nodal efficiency in the right inferior occipital (p = 0.0046) and right postcentral gyri (p = 0.0085), and a 50-node, 73-edge hypoconnectivity sub-network (pFWE = 0.0408); none of these regional findings survived FDR correction. Right inferior occipital efficiency was inversely associated with reminiscence and rumination, and right postcentral efficiency positively with cognition and extraversion. The hippocampal finding was reproduced using KLDs (p = 0.0012) and 4-mm smoothing (p = 0.0049). In 74 patients with major depressive disorder, global topology was likewise preserved but without corresponding nodal alterations.

CONCLUSIONS: Late-life StD shows preserved global topology with focal nodal and sub-network disruptions. Hippocampal hyper-connectivity may reflect compensatory reorganization, whereas occipitotemporal and somatosensory hypoconnectivity is associated with reminiscence, rumination and cognitive function. Individual-level networks may provide a sensitive framework for characterising subclinical network alterations, pending confirmation in larger samples.

PMID:43026081 | DOI:10.1016/j.inpsyc.2026.100307

Document this CPD