Annals of Neurology最新文献

Paramagnetic-rim lesions are a novel diagnostic marker in multiple sclerosis (MS) and are associated with poor prognosis due to their link with chronic inflammation and disease progression. Analyzing 46 postmortem MS cases, researchers found no iron rims in 67 white matter and 85 grey matter spinal cord lesions, despite most being active. In contrast, iron rims appeared in 20.9% of cortical and 80% of subcortical brain lesions, especially in deeper myelin-rich cortical layers. These findings highlight the regional variability of iron accumulation and have important implications for interpreting iron-rims in MS diagnosis, monitoring, and prognostication. ANN NEUROL 2026;99:730-736.

Objective: To assess real-world effectiveness of switching disease-modifying therapy (DMT) in pediatric multiple sclerosis (MS) and clinically isolated syndrome (CIS) initially treated with platform injectables on disease activity.

Methods: Of 2615 pediatric-onset demyelinating disease patients at 12 clinics in the United States (US) Network of Pediatric MS Centers, those with MS/CIS on initial therapy with a platform injectable who switched to another class of platform injectable, oral or infusion DMT were analyzed. Relapse rate was modeled with negative binomial regression, adjusted for preidentified confounders.

Results: A total of 212 children switched DMT before age 18 (67% female, 95% MS). Ninety-three switched from injectable to injectable, 76 injectable to oral, and 43 injectable to infusion. Switchers to oral or infusion were older at onset (injectable 12.3 years, oral 13.5 years, and infusion 14.2 years) and switch (injectable 14.6 years, oral 16.0 years, and infusion 15.7 years). Switchers to infusion DMT were more likely to have enhancing lesions (injectable 45%, oral 28%, and infusion 67%). Compared to injectable (annualized relapse rate [ARR] = 0.88, 95% confidence interval [CI] = 0.52-1.48), relapse rates were lower for injectable to oral (ARR = 0.34, 95% CI = 0.20-0.57; rate ratio: 0.38, 95% CI = 0.21-0.69) and injectable to infusion (ARR = 0.18, 95% CI = 0.09-0.37; rate ratio: 0.21, 95% CI = 0.10-0.44) (p < 0.001). Adjusted number needed to treat in person-years to prevent 1 relapse with oral over injectable was 1.84 (95% CI = 1.03-8.69) and infusion over injectable 1.43 (95% CI = 1.00-3.88).

Interpretation: Switching from platform injectable to oral or infusion compared to other platform injectable DMT led to better disease control in pediatric MS. Long-term safety data are required. ANN NEUROL 2026;99:715-729.

Objective: Frontotemporal lobar degenerations (FTLD)-TDP type C (TDP-C) is distinguished from other FTLD-TDP subtypes by 3 unique features: (1) invariable onset in the anterior temporal lobe (ATL), (2) phosphorylated TDP-43 (pTDP) neurites in cortex, and (3) colocalization of all pTDP deposits with annexin A11 (ANXA11). This article provides a whole-brain anatomical account of TDP-C disease progression in relation to clinical, imaging, and neuropathologic patterns.

Methods: Thirty-two cases with TDP-C were studied, including neuropathologic findings and longitudinal magnetic resonance imaging. In 6 of these cases, cortical and subcortical areas were analyzed using whole hemisphere sections. Five control cases were used for comparison.

Results: Progression was reconstructed with the assumption that regions displaying more severe neurodegeneration at postmortem represented earlier onset sites, an assumption supported by longitudinal imaging. Four stages of cortical TDP-C neuropathology were identified. Subcortically, the fascia dentata contained dense round bodies of pTDP, which did not seem to cause neuronal loss. These deposits were also seen in the nucleus accumbens, islands of Calleja, and the nucleus of the stria terminalis. Cortical predilection sites had higher densities of neurites especially in cases without much tissue destruction. However, cases with more severe overall neurodegeneration displayed a paradoxical scarcity of pTDP deposits at predilection sites. Neurodegeneration patterns identified in whole-hemisphere cases were corroborated by 26 additional cases processed for neuropathologic diagnosis. All pTDP inclusions contained ANXA11. In control cases, the distribution of ANXA11-rich neurons was largely concordant with the pattern of susceptibility to TDP-C.

Interpretation: Neurodegeneration originates in upper cortical layers of ATL and spreads posteriorly along paralimbic mediodorsal and associative ventrolateral pathways. In normal cortex, nuclear TDP-43 is distributed throughout the cortex, whereas normal ANXA11 shows variations that largely mirror TDP-C predilection patterns. The possibility is raised that ANXA11 may be a factor in determining the distribution of TDP-C neurodegeneration. ANN NEUROL 2026;99:639-655.

Objective: Amyotrophic lateral sclerosis (ALS) has a very specific neuroimaging signature, but the molecular underpinnings of the strikingly selective anatomic involvement have not elucidated to date. Accordingly, a large neuroimaging study was conducted with 258 participants to evaluate associations between patterns of neurodegeneration and focal metabolic metrics.

Methods: Structural and diffusivity alterations were systematically evaluated in a genetically stratified cohort. Voxelwise associations between neurodegeneration and physiological mitochondrial indices were systematically evaluated over the entire brain and also examined in specific regions.

Results: Significant topological associations were identified between physiological mitochondria tissue density, nicotinamide adenine dinucleotide (NADH)-ubiquinone oxidoreductase, succinate dehydrogenase, cytochrome c oxidase (COX), mitochondrial respiratory capacity (MRC), tissue respiratory capacity (TRC), and propensity to focal atrophy in ALS. Anatomic correlations between mitochondrial metrics and morphometric change were particularly strong in GGGGCC hexanucleotide repeat carriers in C9orf72. Diffusivity analyses also confirmed associations between brain metabolism and microstructural degeneration. Higher focal mitochondria tissue density was associated with higher likelihood of frontal, temporal, cerebellar, opercular, thalamic, cingulum, putamen, corpus callosum, and corona radiata degeneration. Uncinate fasciculus degeneration was associated with higher Complex I, II, COX, and TRC activity. Topological associations were readily replicated in an external validation cohort.

Interpretation: Our data indicate that brain regions with high metabolic activity are particularly vulnerable to neurodegeneration in ALS. Anatomic associations between physiological cerebral metabolism and patterns of neurodegeneration implicate mitochondrial dysfunction in the pathophysiology of ALS. Although mitochondrial dysfunction may not be the primary etiological factor, it may represent a shared bottleneck of multiple converging molecular and genetic pathways, offering a potential opportunity for meaningful pharmacological intervention. ANN NEUROL 2026;99:614-628.

Objective: Vanishing white matter is a leukodystrophy with remarkable regional variation in disease severity. The cerebral and cerebellar white matter chronically degenerates, while stress-induced episodes of rapid neurological deterioration coincide with the appearance of acute focal lesions in the deep gray structures and brainstem. With clinical recovery, these acute lesions spontaneously improve or completely resolve on magnetic resonance imaging, suggesting endogenous repair. Repair is not observed in cerebral and cerebellar white matter. Here, we investigated whether differences in cellular pathology might explain the regional variation in repair potential.

Methods: We assessed the pathology of astrocytes, microglia, oligodendrocytes, myelin and axons in postmortem human brain tissue of vanishing white matter patients, who died either during or long after episodes of clinical decline, and of controls. We compared the most severely affected frontal white matter with the least affected pons white matter.

Results: In the frontal white matter, astrocytes, microglia and oligodendrocytes were abnormal with evident failure of their mature functions, whereas in the pons white matter, their morphology was relatively preserved with evidence of better preserved functions, suggesting that differences in neuropathology underlie the divergent regional potential for intrinsic repair.

Interpretation: Loss of glial cell functions in the cerebral white matter relates to irreversible degeneration, whereas preserved glial functions and valid inflammatory response in the pons coincide with spontaneous repair in vanishing white matter. Unraveling the cellular behavior underlying the potential for repair in the pons could pave the way to novel treatment strategies for this still incurable disease. ANN NEUROL 2026;99:737-747.

Among patients with acute ischemic stroke achieving successful large vessel recanalization (defined as expanded Thrombolysis in Cerebral Infarction [eTICI ≥2b]), incomplete tissue-level reperfusion, distinct from visually identifiable distal occlusion on digital-subtraction angiography, remains a significant challenge. Persistent tissue-level hypoperfusion, identified on post-thrombectomy perfusion imaging, involves complex pathophysiology comprising 2 primary mechanisms: territorial hypoperfusion, stemming from distal emboli that often manifest after eTICI 2b to 2c or a posteriori with eTICI 3 recanalization, obstructing small distal arterial branches and forming wedge-shaped patterns; and hypoperfusion within the ischemic core, potentially representing capillary no-reflow, a phenomenon of microvascular perfusion failure despite successful macrovascular recanalization well described in the preclinical literature. Such pathophysiological differences have driven inconsistent designations, causing reported incomplete tissue-level reperfusion rate to vary widely (0-42.5%) even in angiographically complete (eTICI 3) recanalization. Further clouding the scene, territorial hypoperfusion may spontaneously reverse within 24 hours (termed "delayed reperfusion"), yet a distinct delayed hypoperfusion can affect necrotic tissue regardless of no-reflow. Whereas persistent hypoperfusion is significantly associated with functional outcomes, the functional impact of true no-reflow remains unclear. Recent positive randomized controlled trials (RCTs) of intra-arterial thrombolysis after successful recanalization offer a promising therapeutic strategy. However, these trials lacked perfusion imaging, whereas the larger functional benefit in patients achieving eTICI 2b suggests intra-arterial thrombolysis likely acted on distal emboli rather than true microvascular dysfunction. Regarding microvascular dysfunction, preclinical findings highlight potential therapeutic strategies such as targeting pericyte constriction or inflammatory responses, that warrant clinical translation. This review synthesizes current evidence on the mechanisms, assessment methods, and therapeutic strategies for addressing incomplete tissue-level reperfusion following thrombectomy. Further research is warranted to establish standardized definitions, develop targeted therapies for both territorial hypoperfusion and true no-reflow, and translate promising preclinical findings into effective clinical interventions. ANN NEUROL 2026;99:668-683.

Objective: Antiseizure medication (ASM) use during pregnancy has increased over the past decade. However, evidence linking prenatal ASM exposure to neurodevelopmental disorders (NDDs) in offspring remains inconsistent. This study evaluated whether prenatal ASM exposure increases the risk of NDDs in children.

Methods: We analyzed data from 5 population-based cohorts of live-born children in Canada (Alberta, Manitoba, Ontario, Quebec; the Canadian Mother-Child Cohort [CAMCCO] cohorts) and the United States (AM-PREGNANT cohort). ASM exposure was defined as maternal prescription fills overlapping the 60 days before birth. NDDs were identified using validated algorithm based on the International Classification of Disease-9/10 codes from inpatient and outpatient records. Within each cohort, Cox proportional hazards models were applied, with adjustment performed separately using (1) covariates and (2) propensity scores. Pooled estimates were obtained using random-effects meta-analysis.

Results: Of 2,910,206 children, 0.47% were exposed to ASMs in the 60 days before birth. Prenatal ASM exposure was associated with a 29% increased risk of NDDs (pooled-adjusted hazard ratio [p-aHR], 1.29; 95% CI: 1.22-1.37; 1,805 exposed cases). In the Canadian cohorts, risks of combined NDDs varied by medication: carbamazepine (p-aHR: 1.50; 95% CI: 1.20-1.87; 262 exposed cases), clonazepam (p-aHR 1.22; 95% CI: 1.12-1.33; 585 exposed cases), topiramate (p-aHR 1.56; 95% CI: 1.04-2.34; 69 exposed cases), and valproic acid (p-aHR 1.38; 95% CI: 1.16-1.65; 134 exposed cases). Although point estimates were higher for polytherapy than monotherapy, the difference was not statistically significant.

Interpretation: Prenatal exposure to certain ASMs was consistently associated with increased risks of NDDs in offspring. These findings support careful, individualized decision-making regarding prenatal ASM use to minimize neurodevelopmental risks. ANN NEUROL 2026;99:761-776.

Objective: Heightened sensitivity to noxious stimulation is a hallmark of chronic pain. Emerging evidence suggests heightened unpleasantness to non-noxious (eg, auditory) aversive stimulation also characterizes chronic pain, but its magnitude, neural mechanisms, and treatment modifiability remain unknown.

Methods: We compared behavioral and neural responses to auditory and pressure stimulation in 142 adults with chronic back pain (CBP) relative to 51 pain-free controls. CBP patients then entered a randomized trial of pain reprocessing therapy (PRT) versus placebo and usual care. During functional magnetic resonance imaging, participants experienced low- and high-intensity aversive sounds and mechanical pressure and provided unpleasantness ratings. Univariate analyses examined responses in primary sensory, sensory-integrative, and midline default mode network regions. Multivariate analyses tested 4 a priori whole-brain patterns, including patterns predictive of fibromyalgia.

Results: CBP patients versus healthy controls reported heightened unpleasantness to auditory stimuli (Hedges' g = 0.95-1.03; p < 0.001) and mechanical pressure (g = 0.49-0.66; p < 0.001). For patients versus controls, auditory stimulation revealed hyperresponsivity in primary auditory cortex and insula, hyporesponsivity in the precuneus and medial prefrontal cortex (g = 0.33-0.59, p < 0.05), and increased expression of generalized and auditory-specific aversive processing patterns (g = 0.33-0.39, p < 0.05) and of fibromyalgia-derived multisensory sensitivity patterns (g = 0.43-0.50, p < 0.01). Longitudinal analysis found that PRT versus placebo led to reduced unpleasantness of low-intensity auditory stimulation, along with increased medial prefrontal cortex responses for PRT versus usual care.

Interpretation: CBP is associated with pronounced auditory hyperresponsivity via modality-specific and modality-general neural pathways, and brain mechanisms overlap with fibromyalgia. PRT versus control produced small reductions in this hyperresponsivity, suggesting potential for PRT to yield broader "central desensitization". ANN NEUROL 2026.