TY - JOUR
T1 - Decreased functional connectivity in post-COVID syndrome patients with high neuroinflammatory activity
AU - Visser, Denise
AU - Pieperhoff, Leo
AU - Lorenzini, Luigi
AU - Golla, Sandeep S.V.
AU - Verfaillie, Sander C.J.
AU - Verveen, Anouk
AU - van de Giessen, Elsmarieke
AU - den Hollander, Marijke E.
AU - Horn, Janneke
AU - van Heugten, Caroline M.
AU - de Jong, Menno D.
AU - van den Wijngaard, Cees C.
AU - van der Maaden, Tessa
AU - Prins, Maria
AU - Schober, Patrick
AU - Schuit, Robert C.
AU - Kassiou, Michael
AU - Windhorst, Albert D.
AU - Appelman, Brent
AU - van Vugt, Michele
AU - van Berckel, Bart N.M.
AU - Knoop, Hans
AU - Wink, Alle Meije
AU - Boellaard, Ronald
AU - Barkhof, Frederik
AU - Tolboom, Nelleke
N1 - Publisher Copyright:
Copyright © 2026. Published by Elsevier Inc.
PY - 2026/6
Y1 - 2026/6
N2 - Introduction Previous research suggests some post-COVID patients with neurocognitive complaints (NCC) show neuroinflammation. Like in inflammatory diseases, this may affect network connectivity. This study aimed to compare resting-state functional connectivity in individuals with and without i) neuroinflammation and ii) persistent post-COVID NCC. Methods Forty-five participants (mean age 49±9 years, 60% female) who had a SARS-CoV-2 infection 27±9 months earlier completed neurocognitive assessment (Checklist-Individual-Strength), 60-minute dynamic [18F]DPA-714 PET scan with arterial sampling for neuroinflammation, and 3T MRI scan for resting-state functional connectivity. Twenty resting-state networks (RSNs) were identified using independent component analysis. Group differences in within-RSN connectivity were analyzed using general linear models. Differences in subcortico-cortical between-RSN connectivity—between brainstem or thalamus and cortical RSNs—were assessed using interaction models. Results Ten participants (22%) showed neuroinflammation on [18F]DPA-714 PET, and 31 (69%) reported persistent NCC. Lower within-RSN connectivity was observed in the visual-peripheral (Nvoxels =3151, PFWE '0.05, tmean=1.77 and tmax=3.34) and dorsal-attention networks (Nvoxels =29, PFWE '0.05, tmean=1.59 and tmax=3.66) in those with neuroinflammation, and in the visual-peripheral (Nvoxels =721, PFWE '0.05, tmean=1.89 and tmax=3.45) and default mode networks (Nvoxels =123, PFWE '0.05, tmean=2.24 and tmax=4.68) in those with NCC. A significant interaction effect showed reduced functional connectivity in a large, bilateral cerebellar cluster (Nvoxels =2648, PFWE '0.05, tmean=2.31 and tmax=3.84) with increasing global [18F]DPA-714 binding in participants with NCC. Lastly, thalamic-somatomotor and brainstem-control network connectivity (between-RSN) was altered in both individuals with persistent NCC and those with neuroinflammation, with thalamic-somatomotor changes mainly driven by NCC and brainstem-control changes by neuroinflammation. Conclusion Our results suggest that neuroinflammation in individuals with persistent NCC after SARS-CoV-2 infection is linked to altered functional connectivity in RSNs central to higher-order cognitive functions.
AB - Introduction Previous research suggests some post-COVID patients with neurocognitive complaints (NCC) show neuroinflammation. Like in inflammatory diseases, this may affect network connectivity. This study aimed to compare resting-state functional connectivity in individuals with and without i) neuroinflammation and ii) persistent post-COVID NCC. Methods Forty-five participants (mean age 49±9 years, 60% female) who had a SARS-CoV-2 infection 27±9 months earlier completed neurocognitive assessment (Checklist-Individual-Strength), 60-minute dynamic [18F]DPA-714 PET scan with arterial sampling for neuroinflammation, and 3T MRI scan for resting-state functional connectivity. Twenty resting-state networks (RSNs) were identified using independent component analysis. Group differences in within-RSN connectivity were analyzed using general linear models. Differences in subcortico-cortical between-RSN connectivity—between brainstem or thalamus and cortical RSNs—were assessed using interaction models. Results Ten participants (22%) showed neuroinflammation on [18F]DPA-714 PET, and 31 (69%) reported persistent NCC. Lower within-RSN connectivity was observed in the visual-peripheral (Nvoxels =3151, PFWE '0.05, tmean=1.77 and tmax=3.34) and dorsal-attention networks (Nvoxels =29, PFWE '0.05, tmean=1.59 and tmax=3.66) in those with neuroinflammation, and in the visual-peripheral (Nvoxels =721, PFWE '0.05, tmean=1.89 and tmax=3.45) and default mode networks (Nvoxels =123, PFWE '0.05, tmean=2.24 and tmax=4.68) in those with NCC. A significant interaction effect showed reduced functional connectivity in a large, bilateral cerebellar cluster (Nvoxels =2648, PFWE '0.05, tmean=2.31 and tmax=3.84) with increasing global [18F]DPA-714 binding in participants with NCC. Lastly, thalamic-somatomotor and brainstem-control network connectivity (between-RSN) was altered in both individuals with persistent NCC and those with neuroinflammation, with thalamic-somatomotor changes mainly driven by NCC and brainstem-control changes by neuroinflammation. Conclusion Our results suggest that neuroinflammation in individuals with persistent NCC after SARS-CoV-2 infection is linked to altered functional connectivity in RSNs central to higher-order cognitive functions.
KW - Neuroinflammation
KW - Post-COVID
KW - Resting-state functional connectivity
KW - rs-fMRI
KW - TSPO PET
UR - https://www.scopus.com/pages/publications/105037829515
U2 - 10.1016/j.neuroimage.2026.121951
DO - 10.1016/j.neuroimage.2026.121951
M3 - Article
C2 - 42034234
AN - SCOPUS:105037829515
SN - 1053-8119
VL - 333
JO - NeuroImage
JF - NeuroImage
M1 - 121951
ER -