Multivariate prediction of long COVID headache in adolescents using gray matter structural MRI features

Abstract

ObjectiveHeadache is among the most frequent symptoms after coronavirus disease 2019 (COVID-19), so-called long COVID syndrome. Although distinct brain changes have been reported in patients with long COVID, such reported brain changes have not been used for predictions and interpretations in a multivariate manner. In this study, we applied machine learning to assess whether individual adolescents with long COVID can be accurately distinguished from those with primary headaches.MethodsTwenty-three adolescents with long COVID headaches with the persistence of headache for at least 3 months and 23 age- and sex-matched adolescents with primary headaches (migraine, new daily persistent headache, and tension-type headache) were enrolled. Multivoxel pattern analysis (MVPA) was applied for disorder-specific predictions of headache etiology based on individual brain structural MRI. In addition, connectome-based predictive modeling (CPM) was also performed using a structural covariance network.ResultsMVPA correctly classified long COVID patients from primary headache patients, with an area under the curve of 0.73 (accuracy = 63.4%; permutation p = 0.001). The discriminating GM patterns exhibited lower classification weights for long COVID in the orbitofrontal and medial temporal lobes. The CPM using the structural covariance network achieved an area under the curve of 0.81 (accuracy = 69.5%; permutation p = 0.005). The edges that classified long COVID patients from primary headache were mainly comprising thalamic connections.ConclusionThe results suggest the potential value of structural MRI-based features for classifying long COVID headaches from primary headaches. The identified features suggest that the distinct gray matter changes in the orbitofrontal and medial temporal lobes occurring after COVID, as well as altered thalamic connectivity, are predictive of headache etiology.

New Daily Persistent Headache in the Pediatric and Adolescent Population: An Updated Review

“…A 2022 study comparing pediatric NDPH patients to healthy controls found decreased cortical thickness and altered resting-state functional connectivity in areas involved in cognitive inhibitory control and pain-related processes [ 10 ]. Interestingly, a 2023 study used structural MRI to distinguish between adolescents with long COVID headache and those with primary headache disorders, including NDPH, revealing potential biomarkers for conditions that, so far, have only clinical criteria [ 32 ]. While a 2022 adult study did not show any gray matter differences in NDPH compared to healthy controls [ 33 ], a 2023 study investigating patients aged 14–70 showed changes in brain morphology such as cortical surface area, cortical thickness, and GM volume, accompanied by abnormal cortical neural activity [ 34 ].…”

“…While many of the other pediatric-focused studies looked at emerging treatment options, they were limited by sample size, as the studies involved relatively few patients and the numbers of NDPH patients within that sample was, with one exception [ 13 ], within the single to low-double digit range [ 9 , 11 , 45 , 46 ]. Likewise, while the studies showing potential biomarkers on MRI, functional connectivity, and brain mapping represent a fascinating and exciting new avenue in diagnosis, the sample sizes were still very low [ 32 , 34 , 35 ]. Further studies are needed to provide new targets for treatment and diagnosis, as well as outcome measures that may more accurately reflect meaningful changes in patient quality of life.…”

“…Within the CDH population at tertiary centers, the prevalence of NDPH in children and adolescents ranges from 0.9 to 35%, compared to 1.7 to 10.8% in adults [19]. It maintains a female predominance throughout life, with a 2.5:1 female/male ratio in adults and a 1.8:1 female/male ratio in children and adolescents, and has an earlier age at onset in females (16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35) than in males (26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45) [2].…”

Bidirectional connectivity alterations in schizophrenia: a multivariate, machine-learning approach

“…To predict schizophrenia using rsFC, we used connectome-based predictive modeling (CPM), a data-driven protocol for developing predictive models of brain–behavior relationships ( 12 ). We modified the CPM approach by replacing its core learning algorithm with a linear support vector machine (SVM) ( 28 , 29 ). We briefly explain how the CPM-SVM procedure works.…”