Brain communications最新文献

{"title":"Evidence that minocycline treatment confounds the interpretation of neurofilament as a biomarker.","authors":"Juliana E Gentile, Christina N Heiss, Taylor L Corridon, Meredith A Mortberg, Stefanie Fruhwürth, Kenia Guzman, Lana Grötschel, Laia Montoliu-Gaya, Kwan Chan, Neil C Herring, Timothy Janicki, Rajaa Nhass, Janani Manavala Sarathy, Brian Erickson, Ryan Kunz, Alison Erickson, Craig Braun, Katherine T Henry, Lynn Bry, Steven E Arnold, Eric Vallabh Minikel, Henrik Zetterberg, Sonia M Vallabh","doi":"10.1093/braincomms/fcaf175","DOIUrl":"10.1093/braincomms/fcaf175","url":null,"abstract":"<p><p>Neurofilament light (NfL) concentration in CSF and blood serves as an important biomarker in neurology drug development. Changes in NfL are generally assumed to reflect changes in neuronal damage, while little is known about the clearance of NfL from biofluids. In a study of asymptomatic individuals at risk for prion disease, both blood and CSF NfL spiked in one participant following a 6-week course of minocycline, absent any other biomarker changes and without subsequent onset of symptoms. We subsequently observed high NfL after minocycline treatment in discarded clinical plasma samples from inpatients, in mouse plasma and in conditioned media from neuron-microglia co-cultures. The specificity and kinetics of NfL response lead us to hypothesize that minocycline does not cause or exacerbate neuronal damage, but instead affects NfL by inhibiting its clearance, posing a potential confounder for the interpretation of this important biomarker.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf175"},"PeriodicalIF":4.1,"publicationDate":"2025-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12100619/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144144980","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sensitivity of diffusion tensor imaging to regional mixed cerebrovascular pathology.","authors":"Jennifer K Ferris, Julia Dahlby, Shie Rinat, Brian Greeley, Joel Ramirez, Sandra E Black, Lara A Boyd","doi":"10.1093/braincomms/fcaf193","DOIUrl":"https://doi.org/10.1093/braincomms/fcaf193","url":null,"abstract":"<p><p>Diffusion tensor imaging is a candidate biomarker in cerebrovascular disease. Yet, little is known about the sensitivity of diffusion tensor imaging to mixed forms of cerebrovascular pathology: stroke and white matter hyperintensities. We evaluated the sensitivity of diffusion tensor imaging to regional lesion load, considering both stroke and white matter hyperintensity lesions. 65 older adults and 39 individuals with chronic stroke underwent diffusion tensor imaging and comprehensive cerebrovascular lesion segmentation. We tested relationships between fractional anisotropy or mean diffusivity and cerebrovascular lesions with linear mixed effects regression. In older adults, tract microstructure related to white matter hyperintensity lesion load (fractional anisotropy: <i>b</i> = -0.003, <i>P</i> = 0.003; mean diffusivity: <i>b</i> = 0.071 × 10^-4, <i>P</i> < 0.001). In individuals with chronic stroke, tract microstructure related to stroke lesion load (fractional anisotropy: <i>b</i> = -0.041, <i>P</i> < 0.001; mean diffusivity: <i>b</i> = 1.460 × 10^-4, <i>P</i> < 0.001), with a significant interaction between stroke and white matter hyperintensity lesion load (fractional anisotropy: <i>b</i> = 0.019, <i>P</i> < 0.001; mean diffusivity: <i>b</i> = -0.727 × 10^-4, <i>P</i> < 0.001). Among both groups, whole-brain normal appearing white matter microstructure did not relate to whole-brain lesion volumes. Our findings provide foundational evidence for the use and interpretation of diffusion tensor imaging as a biomarker in cerebrovascular disease.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf193"},"PeriodicalIF":4.1,"publicationDate":"2025-05-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12120440/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144181557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Delineating network integration and segregation in the pathophysiology of functional neurological disorder.","authors":"Christiana Westlin, Andrew J Guthrie, Cristina Bleier, Sara A Finkelstein, Julie Maggio, Jessica Ranford, Julie MacLean, Ellen Godena, Daniel Millstein, Sara Paredes-Echeverri, Jennifer Freeburn, Caitlin Adams, Christopher D Stephen, Ibai Diez, David L Perez","doi":"10.1093/braincomms/fcaf195","DOIUrl":"10.1093/braincomms/fcaf195","url":null,"abstract":"&lt;p&gt;&lt;p&gt;Functional neurological disorder (FND) is a neuropsychiatric condition that is framed as a multi-network brain problem. Despite this conceptualization, studies have generally focused on specific regions or connectivity features, under-characterizing the complex and nuanced role of resting-state networks in FND pathophysiology. This study employed three complementary graph theory analyses to delineate the functional network architecture in FND. Specifically, we investigated whole-brain weighted-degree, isocortical integration and isocortical segregation extracted from resting-state functional MRI data prospectively collected from 178 participants: 61 individuals with mixed FND; 58 psychiatric controls matched on age, sex, depression, anxiety and post-traumatic stress disorder severity; and 59 age- and sex-matched healthy controls. All analyses were adjusted for age, sex and antidepressant use and focused on differences between FND versus psychiatric controls, with individual-subject maps normalized to healthy controls. Compared to psychiatric controls, patients with mixed FND exhibited increased weighted-degree in the right dorsal anterior cingulate and superior frontal gyrus and the left inferior frontal gyrus and supplementary motor area. Isocortical integration analyses revealed increased &lt;i&gt;between-network&lt;/i&gt; connectivity for somatomotor network areas, with widespread heightened connections to regions of the default mode, frontoparietal and salience networks. Isocortical segregation analyses revealed increased &lt;i&gt;within-network&lt;/i&gt; connectivity for the frontoparietal network. Secondary analyses of functional motor disorder (&lt;i&gt;n&lt;/i&gt; = 46) and functional seizure (&lt;i&gt;n&lt;/i&gt; = 23) subtypes (versus psychiatric controls) revealed both shared and unique patterns of altered connectivity across subtypes, including increased weighted-degree and integrated connectivity in the left posterior insula and anterior/mid-cingulate in functional motor disorder and increased segregated connectivity in the right angular gyrus for functional seizures. In &lt;i&gt;post hoc&lt;/i&gt; between-group analyses, findings remained significant adjusting for depression, anxiety and post-traumatic stress disorder severity, as well as for childhood maltreatment. &lt;i&gt;Post hoc&lt;/i&gt; correlations revealed significant relationships between connectivity metrics in several of these regions and somatic symptom severity across FND and psychiatric control participants. Notably, individual connectivity values were predominantly within the range of healthy controls (with patients with FND generally showing tendencies for increased connectivity and psychiatric controls showing tendencies towards decreased connectivity), indicating subtle shifts in the network architecture rather than gross abnormalities. This study provides novel mechanistic insights (i.e. increased somatomotor integration) and specificity regarding the neurobiology of FND, highlighting both shared mechanisms across subtypes and ","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf195"},"PeriodicalIF":4.1,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12107243/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144164291","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Plasma α-synuclein domain profiles across α-synucleinopathies.","authors":"Marie-Laure Pons, Pablo Mohaupt, Jérôme Vialaret, Etienne Mondesert, Margaux Vignon, Salomé Coppens, Moreau Stéphane, Sylvain Lehmann, Christophe Hirtz","doi":"10.1093/braincomms/fcaf189","DOIUrl":"10.1093/braincomms/fcaf189","url":null,"abstract":"<p><p>The differential diagnosis of α-synucleinopathies, including Parkinson's disease, dementia with Lewy bodies (DLB) and multiple system atrophy (MSA), remains challenging due to overlapping clinical features and the absence of reliable biomarkers. We developed a targeted mass spectrometry assay to profile α-synuclein peptides in plasma from Parkinson's disease (<i>n</i> = 82), DLB (<i>n</i> = 32), MSA (<i>n</i> = 8) and controls (<i>n</i> = 21). We hypothesized that disease-specific truncations or post-translational modifications would alter levels of non-modified α-synuclein peptides across α-synucleinopathies. The assay quantified non-modified peptides derived from the N-terminus and non-amyloid component (NAC) domain, regions implicated in aggregate formation. Although peptide levels were consistent across disease groups, a distinct NAC domain pattern observed in MSA may reflect unique pathological processes. This study presents the first blood-based profiling of α-synuclein peptides in these disorders, offering a basis for further investigation into disease mechanisms. Refinement of the assay to include post-translational modifications could enhance understanding of α-synucleinopathies and support future biomarker development.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf189"},"PeriodicalIF":4.1,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12107063/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144164293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ending nuclear weapons, before they end us ^†.","authors":"Kamran Abbasi, Parveen Ali, Virginia Barbour, Marion Birch, Inga Blum, Peter Doherty, Andy Haines, Ira Helfand, Richard Horton, Kati Juva, Jose F Lapena, Robert Mash, Olga Mironova, Arun Mitra, Carlos Monteiro, Elena N Naumova, David Onazi, Tilman Ruff, Peush Sahni, James Tumwine, Carlos Umaña, Paul Yonga, Chris Zielinski","doi":"10.1093/braincomms/fcaf154","DOIUrl":"10.1093/braincomms/fcaf154","url":null,"abstract":"","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf154"},"PeriodicalIF":4.1,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12086005/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144103309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thalamic disconnection from prefrontal cognitive control networks contributes to thalamic aphasia.","authors":"Anika Stockert, Sophia Hormig-Rauber, Julian Klingbeil, Sophie Marie Meixensberger, Karl-Titus Hoffmann, Dorothee Saur, Max Wawrzyniak","doi":"10.1093/braincomms/fcaf191","DOIUrl":"10.1093/braincomms/fcaf191","url":null,"abstract":"<p><p>Language impairments after thalamic lesions, referred to as thalamic aphasia, underscore a subcortical involvement in language processing. In this study, we investigated how the thalamus structurally connects to the cortex to support language functions. Our hypothesis posits that disconnection of white matter tracts between the left thalamus and regions of left hemisphere language and cognitive control networks, such as prefrontal, inferior frontal, and temporal cortices, are associated with thalamic aphasia. We employed a non-parametric lesion-network mapping approach in a retrospective cohort of patients with first-ever thalamic stroke. This method enables the identification of structural disconnections that disrupt signal transmission along white matter fibre pathways, subsequently impairing processing within brain networks. To investigate potential associations between disconnection patterns and thalamic aphasia, we individually mapped fibre tracts affected by the thalamic stroke lesions using diffusion-weighted normative structural connectome data. Statistical comparisons were then made between disconnection maps of patients with and without language impairments. The study encompassed 101 patients, with a mean age of 64.1 years (standard deviation, 14.6), including 57 patients with left-sided, 42 with right-sided, and 2 with bilateral thalamic lesions. We observed that language impairments were linked to disconnection of fibres in the left anterior limb of the internal capsule. These fibres constitute a pathway within the anterior thalamic radiation, connecting the mediodorsal thalamus to a region in the left dorsomedial prefrontal cortex. An additional exploratory analysis revealed functional connectivity between this cortical area and the left hemisphere's language-related inferior frontal and lateral temporal cortices. Meanwhile, we found no evidence for direct structural disconnection between the thalamus and left inferior frontal or temporal cortices. Interrupted signalling along the left anterior thalamic radiation to the left dorsomedial prefrontal cortex is a potential mechanism underlying thalamic aphasia. Given its significance for the subcortical involvement in language processing, our findings suggest a plausible path for thalamic signals to participate in the interplay between networks for cognitive control and language processing.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf191"},"PeriodicalIF":4.1,"publicationDate":"2025-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12116879/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144176115","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cannabis- and HIV-related perturbations to the cortical gamma dynamics supporting inhibitory processing.","authors":"Mikki Schantell, Mia C Lulli, Kellen M McDonald, Lucy K Horne, Jason A John, Anna T Coutant, Hannah J Okelberry, Ryan Glesinger, Yasra Arif, Jennifer L O'Neill, Sara H Bares, Pamela E May-Weeks, Tony W Wilson","doi":"10.1093/braincomms/fcaf190","DOIUrl":"10.1093/braincomms/fcaf190","url":null,"abstract":"<p><p>The main psychoactive component in cannabis-Δ⁹-tetrahydrocannabinol-is known to have anti-inflammatory properties and to alter gamma oscillations, pointing to its potential as a therapeutic agent for people with HIV (PWH). However, it remains unknown how cannabis use among PWH interacts with the neural circuitry underlying inhibitory processing. Herein, using a cross-sectional study design, we collected data from 108 cannabis users and non-users with and without HIV. Participants were interviewed regarding their substance use history and completed a paired-pulse somatosensory stimulation paradigm during magnetoencephalography (MEG). MEG data were imaged using a beamformer and peak voxel time series data were extracted to examine neural oscillations in response to the stimulation and the strength of spontaneous activity in the same tissue during the baseline period. Across all participants, we observed robust gamma oscillations following stimulation in the left primary somatosensory cortices, with responses to the second stimulation being strongly attenuated relative to the first, thus demonstrating somatosensory gating. PWH who used cannabis exhibited stronger oscillatory gamma activity compared with non-users with HIV, while the latter group also exhibited elevated spontaneous gamma activity relative to all other groups. Finally, we found that a longer duration of time since HIV diagnosis was associated with less efficient inhibitory processing among PWH who did not use cannabis, but not among PWH who regularly use cannabis. These findings provide new evidence that cannabis use may mitigate the harmful effects of HIV on oscillatory and spontaneous gamma activity serving inhibitory processing.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf190"},"PeriodicalIF":4.1,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12103976/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144153130","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Vasomotion loss precedes impaired cerebrovascular reactivity and microbleeds in cerebral amyloid angiopathy.","authors":"Mariel G Kozberg, Leon P Munting, Lee H Hanlin, Corinne A Auger, Maarten L van den Berg, Baudouin Denis de Senneville, Lydiane Hirschler, Jan M Warnking, Emmanuel L Barbier, Christian T Farrar, Steven M Greenberg, Brian J Bacskai, Susanne J van Veluw","doi":"10.1093/braincomms/fcaf186","DOIUrl":"10.1093/braincomms/fcaf186","url":null,"abstract":"<p><p>Cerebral amyloid angiopathy (CAA) is a cerebral small vessel disease in which amyloid-β accumulates in vessel walls. CAA is a leading cause of symptomatic lobar intracerebral haemorrhage and an important contributor to age-related cognitive decline. Recent work has suggested that vascular dysfunction may precede symptomatic stages of CAA, and that spontaneous slow oscillations in arteriolar diameter (termed vasomotion), important for amyloid-β clearance, may be impaired in CAA. To systematically study the progression of vascular dysfunction in CAA, we used the APP23 mouse model of amyloidosis, which is known to develop spontaneous cerebral microbleeds mimicking human CAA. Using <i>in vivo</i> 2-photon microscopy, we longitudinally imaged unanesthetized APP23 transgenic mice and wildtype (WT) littermates from 7 to 14 months of age, tracking amyloid-β accumulation and vasomotion in individual pial arterioles over time. MRI was used in separate groups of 12-, 18- and 24-month-old APP23 transgenic mice and WT littermates to detect microbleeds and to assess cerebral blood flow (CBF) and cerebrovascular reactivity (CVR) with pseudo-continuous arterial spin labelling. We observed a significant decline in vasomotion with age in APP23 mice, while vasomotion remained unchanged in WT mice with age. This decline corresponded in timing to initial vascular amyloid-β deposition (∼8-10 months of age), although it was more strongly correlated with age than with vascular amyloid-β burden in individual arterioles. Declines in vasomotion preceded the development of MRI-visible microbleeds and the loss of smooth muscle actin in arterioles, both of which were observed in the majority of APP23 mice by 18 months of age. Additionally, CBF and evoked CVR were intact in APP23 mice at 12 months of age, but significantly lower in APP23 mice by 24 months of age. Our findings suggest that a decline in spontaneous vasomotion is an early, potentially pre-symptomatic, manifestation of CAA and vascular dysfunction, and a possible future treatment target.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf186"},"PeriodicalIF":4.1,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12096159/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144129847","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cardiovascular risk factors modulate the effect of brain imaging-derived phenotypes on ischaemic stroke risk.","authors":"Yuan-Yuan Liang, Meng-Jie Li, Dong-Rui Ma, Meng-Nan Guo, Xiao-Yan Hao, Shuang-Jie Li, Chun-Yan Zuo, Chen-Wei Hao, Zhi-Yun Wang, Yan-Mei Feng, Chenyuan Mao, Chan Zhang, Bo Song, Yuming Xu, Changhe Shi","doi":"10.1093/braincomms/fcaf183","DOIUrl":"10.1093/braincomms/fcaf183","url":null,"abstract":"<p><p>Studies have shown that cardiovascular risk factors are closely related to the occurrence of stroke, especially ischaemic stroke, as they can lead to changes in brain structure and function. However, the role of cardiovascular risk factors-induced changes in brain structure and function in the development of ischaemic stroke has not been studied. The aim of this study is thus to explore the causal association among cardiovascular risk factors, brain phenotypes and ischaemic stroke by assessing Mendelian randomization. We used univariate Mendelian randomization to sequentially investigate the causal effects of the 12 most common cardiovascular risk factors on brain structure and 3935 brain imaging-derived phenotypes in the development of ischaemic stroke. We also examined the mediating effect of brain structure on blood pressure-induced ischaemic stroke using a multivariable Mendelian randomization test. We tested the reliability of our results using the Steiger test, heterogeneity test, horizontal pleiotropy test and leave-one-out method. We found that 8 of the 12 examined cardiovascular risk factors were associated with 538 brain imaging-derived phenotypes, and 9 of the 12 cardiovascular risk factors were associated with IS. The main cardiovascular risk factors associated with brain imaging-derived phenotypes and ischaemic stroke was blood pressure (systolic and diastolic), which can affect the occurrence of ischaemic stroke through 6 types of brain imaging-derived phenotypes. However, extrapolation of our findings to other ethnic groups is challenging, and the possibility of reverse causality cannot be completely ruled out. This study identifies the role of cardiovascular risk factors, especially blood pressure, in affecting brain structure and ischaemic stroke risk. The findings assist in early risk detection and enhance stroke prevention strategies, also hinting at non-vascular factors' involvement.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf183"},"PeriodicalIF":4.1,"publicationDate":"2025-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12089767/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144112797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multimodal deep learning model for prediction of prognosis in central nervous system inflammation.","authors":"Bo Kyu Choi, Yoonhyeok Choi, Sooyoung Jang, Woo-Seok Ha, Soomi Cho, Kimoon Chang, Beomseok Sohn, Kyung Min Kim, Yu Rang Park","doi":"10.1093/braincomms/fcaf179","DOIUrl":"10.1093/braincomms/fcaf179","url":null,"abstract":"<p><p>Inflammatory diseases of the CNS impose a substantial disease burden, necessitating prompt and appropriate prognosis prediction. We developed a multimodal deep learning model integrating clinical features and brain MRI data to enhance early prognosis prediction of CNS inflammation. This retrospective study used thin-cut T1-weighted brain MRI scans and the clinical variables of patients with CNS inflammation who were admitted to a tertiary referral hospital between January 2010 and December 2023. Data collected after January 2022 served as the external test set. 3D MRI images were first segmented into 43 brain regions using the FastSurfer library. The segmented images were then processed through a 3D convolutional neural network model for feature extraction and vectorization, after which they were integrated with clinical features for prediction. The performance of each artificial intelligence model was assessed using accuracy, F1 score, area under the receiver operating characteristic curve and area under the precision-recall curve. The internal dataset comprised 413 images from 291 patients (mean age, 45.5 years ± 19.3 [SD]; 151 male patients; 54 with poor prognosis). The external dataset comprised 210 images from 106 patients (mean age, 45.5 years ± 18.9 [SD]; 59 male patients; 31 with poor prognosis). The multimodal deep learning model outperformed unimodal models across all aetiological groups, achieving area under the receiver operating characteristic curve values of 0.8048 for autoimmune, 0.9107 for bacterial, 1.0000 for tuberculosis and 0.9242 for viral infections. Furthermore, artificial intelligence assistance improved clinicians' prognostic accuracy, as demonstrated in comparisons with neurologists, paediatricians and radiologists. Our findings demonstrate that the multimodal deep learning model enhances artificial intelligence-assisted prognosis prediction in CNS inflammation, improving both model performance and clinician decision-making.</p>","PeriodicalId":93915,"journal":{"name":"Brain communications","volume":"7 3","pages":"fcaf179"},"PeriodicalIF":4.1,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12082089/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096185","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}