BACKGROUND:Lysosomal dysfunction is central to Parkinson's disease (PD) pathogenesis, with GBA1 representing the strongest established genetic risk factor. Numerous other genes involved in lysosomal sphingolipid, glycosphingolipid, and ceramide metabolism have been proposed as contributors to PD, highlighting the need for genetic analyses across these pathways. OBJECTIVES:The aim was to evaluate the contribution of rare variants across lysosomal genes to PD risk. METHODS:We analyzed rare variants (minor allele frequency ≤0.01) across 36 lysosomal genes in 8267 individuals with PD and 68,208 controls, including 793 early-onset PD (≤50 years) cases. Targeted sequencing was performed in four cohorts at McGill University (3456 cases and 2664 controls) and combined with whole-genome sequencing data from the United Kingdom (UK) Biobank (2848 cases, 62,451 controls) and the Accelerating Medicines Partnership-PD cohort (1963 cases, 3093 controls). Associations were tested using Sequence Kernel Association Test-Optimal across variant classes (rare variants, nonsynonymous, loss-of-function, and predicted damaging variants with combined annotation-dependent depletion score >20), followed by meta-analysis across cohorts. Domain-level analyses were performed for variants located within protein domains. False discovery rate (FDR) correction was applied. RESULTS:Meta-analysis identified a significant association between rare variants in ST3GAL3 and Parkinson's disease (Pfdr = 0.04). Domain-based analyses showed enrichment of nonsynonymous variants within the β-acetyl-hexosaminidase-like domain of HEXA (P = 8.0 × 10-4), although this signal did not survive correction (Pfdr = 0.154). In early-onset PD, domain-based analyses identified significant associations in NAGLU (Pfdr = 7.3 × 10-6) and ST3GAL5 (Pfdr = 0.03). CONCLUSIONS:Rare variants across multiple lysosomal pathways, particularly those related to sialylation, ganglioside metabolism, ceramide biology, and lysosomal proteolysis, may contribute to PD susceptibility beyond GBA1, highlighting pathways for future replication and investigation. © 2026 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Objective.Transcranial ultrasound stimulation (TUS) is a noninvasive neuromodulation technique offering millimeter-scale precision and deep targeting. However, the trade-off between fundamental frequency, focal size, and efficacy in humans remains unclear. This study aimed to compare the effects of broad versus narrow acoustic focus on corticospinal excitability using 250 kHz and 825 kHz TUS, and to evaluate whether multi-focal stimulation mitigates limitations of narrow beams.Approach.Twenty healthy adults underwent four randomized, double-blind sessions: 250 kHz unifocal, 825 kHz unifocal, 825 kHz multi-focal, and sham. TUS was delivered to the left primary motor cortex using a 128-element phased-array transducer integrated with neuronavigation. Corticospinal excitability was assessed using motor-evoked potentials (MEPs) elicited by transcranial magnetic stimulation at baseline and 5-, 30-, and 60 min post-stimulation. Reaction (RT) time and accuracy were measured using a Go/No-Go task.Main results.Linear mixed-effects modeling revealed a significant inhibition of corticospinal excitability, characterized by decreased MEP amplitudes, for both the 250 kHz unifocal and 825 kHz multi-focal conditions compared to sham (both withp= 0.018). The 825 kHz unifocal stimulation demonstrated only a non-significant trend toward inhibition (p= 0.071). Effects were transient, most evident at 5- and 30 min post-TUS. Behavioral performance (measured via RT and task accuracy) remained unchanged across all conditionsSignificance.Findings suggest that spatial coverage, achieved through broader 250 kHz beams or multi-focal cycling, is critical for effective TUS-induced neuromodulation of the motor cortex. Future studies should optimize dosing and targeting strategies to balance precision and efficacy.

Despite growing understanding of the benefits of having Findable, Accessible, Interoperable, and Reusable (FAIR) data, many datasets still cannot be shared. Federated analysis methods can enable multisite studies that do not require the sharing of participant-level information. However, there are many practical hurdles that prevent the large-scale adoption of federated methods. We discuss challenges related to cross-site data preparation for federated learning, present solutions offered by recent neuroinformatics projects, and showcase an example of tool integration applied to neurodegenerative disease data.

ObjectiveGiven the clinical heterogeneity of Parkinson’s disease (PD), identification of early -stage subgroups with shared non-motor symptom (NMS) profiles may clarify its pathophysiology. This study used latent-profile analyses (LPA) to define subgroups based on sleep disturbances, cognitive performance and neuropsychiatric symptoms, and examined dopaminergic function and brain volume differences between them.MethodsWe analyzed data from 51 cognitively normal non-PD older adults and 105 early-stage PD participants from the iPARK trial, including 19 who underwent [11C]-raclopride PET/MR. Participants completed the Hospital Anxiety and Depression Scale, the short version of the Karolinska Sleep Questionnaire and a battery of neuropsychological tests. LPA were used in PD to identify subgroups based on NMS profiles, which were then characterized and examined in relation to dopaminergic integrity and brain morphology.ResultsLPA identified a two-cluster solution as the best fit. Group 1 (N = 49) showed poorer working memory, executive function and processing speed along with greater daytime sleepiness, depression and anxiety. Group 2 (N = 56) exhibited less affected cognitive function and minimal NMS. Groups were similar in demographics, disease duration, motor symptom severity and medication, but differed on UPDRS-1 NMS. Group 1 demonstrated significantly reduced [11C]-raclopride binding potential compared to Group 2 in the left putamen at both ROI- and voxel-wise analysis.ConclusionThese findings indicate clinically distinct subgroups in early-stage PD. Greater NMS burden is linked to impaired dopaminergic integrity, suggesting a potential neurobiological signature. Early identification of such subgroups may improve understanding of disease heterogeneity and support personalized management and interventions.Clinical trial registrationhttps://clinicaltrials.gov/study/NCT03680170?id=NCT03680170&rank=1, identifier (NCT03680170).

Federated learning (FL) and travelling model (TM) allow privacy-preserving model training across sites without sharing patient-sensitive data. While both approaches have shown success, they face unique challenges related to distribution shifts between sites. To address this, we propose FedTM, a hybrid framework combining the strengths of FL and TM. FedTM begins with FL warmup training at sites with larger datasets, followed by sequential refinement through TM across all sites. We evaluated FedTM for Parkinson's disease classification using 1817 brain scans from 83 international sites. Model performance, misclassification disparities, and communication costs were computed and compared to standard FL and TM approaches. Our results reveal that FedTM improves AUROC from 77 ± 0.01% to 82 ± 0.01%, reduces misclassification disparities from 34 ± 0.01% to 26 ± 0.01%, and decreases training load for smaller sites from 22 to 12 cycles. These advancements mark an important step toward promoting global healthcare equity and advancing responsible AI development.

Background:Variants in GBA1 are important genetic risk factors for synucleinopathies, including Parkinson's disease (PD). While several GBA1 variants are established risk or severity modifiers, the role of the p.E427K variant remains unclear. Objective:To determine whether the GBA1 p.E427K variant is associated with risk of synucleinopathies. Methods:We performed a meta-analysis of case-control studies reporting the frequency of GBA1 p.E427K (p.E388K) in PD and related synucleinopathies. Data were obtained from published studies, open-access resources, and large cohorts, including in-house datasets. Odds ratios (ORs) were calculated for each cohort and pooled using a random-effects model. Results:Across 67,484 patients and 124,079 controls, GBA1 p.E427K was associated with increased disease risk (pooled OR = 1.87, 95% CI 1.28-2.72, P = 0.001). Enzymatic data showed reduced glucocerebrosidase activity in carriers. Conclusions:The GBA1 p.E427K variant is a risk factor for synucleinopathies and should be considered in genetic studies and clinical trials.

ABSTRACT:Noninvasive neuromodulatory techniques provide important means to study the role of brain regions activated by nociceptive stimuli in pain perception. This study investigates the role of the primary somatosensory cortex (S1) and the ventral posterolateral nucleus (VPL) of the thalamus in acute pain perception using transcranial ultrasound stimulation (TUS). Twenty-five healthy participants underwent a double-blind, sham-controlled, within-subject experimental design. Transcranial ultrasound stimulation was applied to the left S1 and left VPL in separate sessions, with quantitative sensory testing performed before and after stimulation. Measures included heat pain threshold (HPT), heat pain tolerance (HPTol), warm detection threshold (WDT), mechanical detection threshold (MDT), and pressure pain threshold (PPT). Stimulation of the left S1 significantly lowered HPT ( P = 0.013) and HPTol ( P = 0.040) on the contralateral hand, with median differences of -0.6°C (IQR [-1.20, 0.35]) and -0.2°C (IQR [-1.00, 0.3]), respectively. In addition, both S1 and VPL stimulation led to bilateral reductions in WDT ( P < 0.001), with median decreases ranging from -0.25°C to -0.35°C (IQRs ranging from -0.95 to 0.30). No significant changes were observed in MDT ( P > 0.89) or PPT ( P > 0.78). These findings suggest the involvement of S1 in pain perception, particularly in modulating heat pain sensitivity. The modulation of warm detection by both S1 and VPL further suggests that TUS can influence sensory processing at multiple levels of the somatosensory pathway. Further research is needed to replicate the present findings, elucidate the underlying biophysical mechanisms, and optimize stimulation protocols for clinical applications.

Abstract Background Variants in GBA1 are important genetic risk factors for synucleinopathies, including Parkinson's disease (PD). Although several GBA1 variants are established risk or severity modifiers, the role of the p.E427K variant remains unclear. Objective The aim was to determine whether the GBA1 p.E427K variant is associated with risk of synucleinopathies. Methods We performed a meta‐analysis of case–control studies reporting the frequency of GBA1 p.E427K (p.E388K) in PD and related synucleinopathies. Data were obtained from published studies, open‐access resources, and large cohorts, including in‐house datasets. Odds ratios (OR) were calculated for each cohort and pooled using a random‐effects model. Results Across 67,221 patients and 123,832 controls, GBA1 p.E427K was associated with increased disease risk (pooled OR = 1.94, 95% confidence interval 1.33–2.84, P = 0.0007). Enzymatic data showed reduced glucocerebrosidase activity in carriers. Conclusions The GBA1 p.E427K variant is a risk factor for synucleinopathies and should be considered in genetic studies and clinical trials. © 2026 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Background and ObjectivesIntensive speech therapy may improve recovery from poststroke aphasia. Further evidence suggests that pairing repetitive transcranial magnetic stimulation (rTMS) with intensive speech therapy might augment outcomes. This sham-controlled randomized clinical trial evaluated the efficacy of 1-Hz rTMS over the right pars triangularis combined with multimodality aphasia therapy (M-MAT) in chronic poststroke aphasia. MethodsA parallel-group, double-blind, sham-controlled randomized clinical trial was conducted between April 2021 and May 2023 at an outpatient neurorehabilitation clinic. Individuals with chronic nonfluent aphasia after left middle cerebral artery stroke (>6 months from stroke) were enrolled and randomly assigned to receive either rTMS or sham stimulation combined with 35 hours of M-MAT over 10 days. The primary outcome was the Western Aphasia Battery aphasia quotient (WAB-AQ) measured at 3 weeks and 15 weeks. Intention-to-treat analysis examined treatment effects over time using linear mixed models. ResultsA total of 44 participants were randomized. Forty-three (mean [SD] age, 63.4 [12.3] years; 14 women [32.6%]) completed the intervention. Overall, WAB-AQ scores improved from baseline to 15 weeks regardless of rTMS allocation (mean difference 5.33, 95% CI 2.9-7.8, p < 0.001). We observed a significant group-by-time interaction (beta = 0.31, p = 0.024), suggesting that those who received rTMS combined with M-MAT improved more over time than those who received sham. At 15 weeks, the rTMS group demonstrated significantly less word-finding difficulties and more complete and longer sentences with fewer pauses compared with sham as indicated by higher WAB-AQ scores (mean difference 4.1 points, 95% CI 0.6-7.6, p = 0.022). The change from baseline at 15 weeks was greater in the rTMS group (7.6 points, 95% CI 4.1-11.1) compared with sham (3.0 points, 95% CI -0.3 to 5.2; mean difference 4.6 points, 95% CI 0.6-8.6, p = 0.024). DiscussionIntensive administration of M-MAT alone improves speech production in patients with chronic poststroke aphasia. Combining 1-Hz rTMS with M-MAT is associated with supplemental improvements in aphasia severity at follow-up. rTMS is a promising candidate as an adjuvant therapy to M-MAT. Trial Registration InformationClinicalTrials.gov Identifier: NCT04102228. Classification of EvidenceThis study provides Class III evidence that in patients with aphasia 6 or more months after a stroke, 1-Hz rTMS combined with intensive M-MAT improves WAB-AQ more than sham stimulation plus M-MAT.

Distributed learning enables collaborative machine learning model training without requiring cross-institutional data sharing, thereby addressing privacy concerns. However, local quality control variability can negatively impact model performance while systematic human visual inspection is time-consuming and may violate the goal of keeping data inaccessible outside acquisition centers. This work proposes a novel self-supervised method to identify and eliminate harmful data during distributed learning model training fully-automatically. Harmful data is defined as samples that, when included in training, increase misdiagnosis rates. The method was tested using neuroimaging data from 83 centers for Parkinson's disease classification with simulated inclusion of a few harmful data samples. The proposed method reliably identified harmful images, with centers providing only harmful datasets being easier to identify than single harmful images within otherwise good datasets. While only evaluated using neuroimaging data, the presented method is application-agnostic and presents a step towards automated quality control in distributed learning.

The presence of non-motor symptoms (NMS) such as olfactive deficit or neuropsychiatric symptoms has been associated with the diagnosis of Parkinson’s Disease (PD). NMS are also associated with different brain structural features underlying distinctive processes in PD. NMS has been poorly studied in patients with a PD-like clinical profile, showing Scans Without Evidence of Dopaminergic Deficit (SWEDD). This study proposes to compare classification models differentiating PD, SWEDD and Healthy Controls (HC) based on NMS and neurostructural factors. 683 participants (382 PD diagnosed in the last 2 years, 48 with SWEDD, 170 HC) from the PPMI dataset were compared based on available assessments. Each participant underwent an olfactive, neuropsychiatric and sleep assessment, and a 3T MRI. Brain volumes were extracted and standardized from each MRI. Classifications were based on logistic regressions using 5-fold cross-validation models combining different NMS and MRI data and determining their involvement in differentiation between patient subgroups (PD vs. SWEDD) or between patients and HC. NMS were significant factors in PD vs. SWEDD, PD vs. HC and SWEDD vs. HC classifiers, when considered alone or in combination with MRI data. No classification models were significantly different from chance based-on MRI, nor more accurate combining NMS and MRI when compared with models based on NMS only. These results highlight the importance of NMS in differentiating between PD and SWEDD, PD and HC, SWEDD and HC. However, classical imaging data such as cortical and subcortical volumetry seems insufficient to improve these classifications. Other imaging features such as connectivity could also be studied.

BACKGROUND:Synucleinopathies include a spectrum of disorders varying in features and severity, including idiopathic/isolated REM sleep behaviour disorder (iRBD), Parkinson's disease (PD), and dementia with Lewy bodies (DLB). Distinct brain atrophy patterns may already be seen in iRBD; however, how brain atrophy begins and progresses remains unclear. METHODS:A multicentric cohort of 1276 participants (451 polysomnography-confirmed iRBD, 142 PD with probable RBD, 87 DLB, and 596 controls) underwent T1-weighted MRI and longitudinal clinical assessments. Brain atrophy was quantified using vertex-based cortical surface reconstruction and volumetric segmentation. The unsupervised machine learning algorithm, Subtype and Stage Inference (SuStaIn), was used to reconstruct spatiotemporal patterns of brain atrophy progression. FINDINGS:SuStaIn identified two distinct subtypes of brain atrophy progression: 1) a "cortical-first" subtype, with atrophy beginning in the frontal lobes and involving the subcortical structures at later stages; and 2) a "subcortical-first" subtype, with atrophy beginning in the limbic areas and involving cortical structures at later stages. Both cortical- and subcortical-first subtypes were associated with a higher rate of increase in MDS-UPDRS-III scores over time, but cognitive decline was subtype-specific, being associated with advancing stages in patients classified as cortical-first but not subcortical-first. Classified patients were more likely to phenoconvert over time compared to stage 0/non-classified patients. Among the 88 patients with iRBD who phenoconverted during follow-up, those classified within the cortical-first subtype had a significantly increased likelihood of developing DLB compared to PD, unlike those classified within the subcortical-first subtype. INTERPRETATION:There are two distinct atrophy progression subtypes in iRBD, with the cortical-first subtype linked to an increased likelihood of developing DLB, while both subtypes were associated with worsening parkinsonian motor features. This underscores the potential utility of subtype identification and staging for monitoring disease progression and patient selection for trials. FUNDING:This study was supported by grants to S.R. from Alzheimer Society Canada (0000000082) and by Parkinson Canada (PPG-2023-0000000122). The work performed in Montreal was supported by the Canadian Institutes of Health Research (CIHR), the Fonds de recherche du Québec - Santé (FRQS), and the W. Garfield Weston Foundation. The work performed in Oxford was funded by Parkinson's UK (J-2101) and the National Institute for Health Research (NIHR) Oxford Biomedical Research Centre (BRC). The work performed in Prague was funded by the Czech Health Research Council (grant NU21-04-00535) and by The National Institute for Neurological Research (project number LX22NPO5107), financed by the European Union - Next Generation EU. The work performed in Newcastle was funded by the NIHR Newcastle BRC based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The work performed in Paris was funded by grants from the Programme d'investissements d'avenir (ANR-10-IAIHU-06), the Paris Institute of Neurosciences - IHU (IAIHU-06), the Agence Nationale de la Recherche (ANR-11-INBS-0006), Électricité de France (Fondation d'Entreprise EDF), the EU Joint Programme-Neurodegenerative Disease Research (JPND) for the Control-PD Project (Cognitive Propagation in Prodromal Parkinson's disease), the Fondation Thérèse et René Planiol, the Fonds Saint-Michel; by unrestricted support for research on Parkinson's disease from Energipole (M. Mallart) and the Société Française de Médecine Esthétique (M. Legrand); and by a grant from the Institut de France to Isabelle Arnulf (for the ALICE Study). The work performed in Sydney was supported by a Dementia Team Grant from the National Health and Medical Research Council (#1095127). The work performed in Cologne was funded by the Else Kröner-Fresenius-Stiftung (grant number 2019_EKES.02), the Köln Fortune Program, Faculty of Medicine, University of Cologne, and the "Netzwerke 2021 Program (Ministry of Culture and Science of Northrhine Westphalia State). The work performed in Aarhus was supported by funding from the Lundbeck Foundation, Parkinsonforeningen (The Danish Parkinson Association), and the Jascha Foundation.

LRRK2 variants are key genetic risk factors for Parkinson’s Disease (PD). We conducted a per-domain rare coding variant burden analysis, including 8,888 PD cases and 69,412 controls. In meta-analysis, the Kinase domain was strongly associated with PD (Exonic: PFDR = 1.61 × 10−22, Non-synonymous: PFDR = 1.54 × 10−23, CADD > 20: PFDR = 3.09 × 10−24). Excluding the p.G2019S variant nullified this effect. Nominal associations were found in the ANK and Roc-COR domains, with potentially protective variants, p.R793M and p.Q1353K.

BACKGROUND:Deep brain stimulation (DBS) is an established treatment for Parkinson's disease (PD) in appropriately selected patients. DBS may be underused in certain patient populations, especially women and racialized groups. Barriers and biases to receiving DBS that could account for underuse among these groups are not well studied in Canada. OBJECTIVE:We aim to better characterize the disparities in gender, ethnicity and other demographic factors among patients referred for and receiving DBS. METHODS:We performed a retrospective chart review and phone survey of DBS patients treated at two Canadian centers and from Canada Open Parkinson Network (C-OPN). Gender, ethnicity, marital status, native language, birth country, urban versus rural residency, level of education, household income and mode of referral were studied. RESULTS:Among all participants, more men than women received DBS. Most patients (81.8-94.1%) in both referral and implanted groups were White. The gender and ethnicity of this cohort do not represent Canadian demographics. Patients referred and receiving surgery had higher educational level compared with general Canadian population. Being married was positively associated with DBS referral and implantation. CONCLUSION:Significant ethnic and gender disparities in receiving DBS exist. Educated White men were overrepresented. Further actions need to be taken to expand the accessibility of this important treatment to all eligible PD patients with an effort to provide equitable care to women, racialized groups and those who cannot advocate for themselves in Canada.

Mild Cognitive Impairment (MCI) may be caused by mixed pathologies. Blood markers indicative of Alzheimer pathology or neurodegeneration have not been extensively explored in patients with MCI who have features of Lewy Body disease (LB-MCI) (cognitive fluctuations, parkinsonism, hallucinations, and REM-sleep behavior disorder (RBD)). We compared plasma levels of amyloid beta (Aß), tau, glial-fibrillary acidic protein (GFAP), and neurofilament light chain (NfL) between participants in COMPASS-ND who met criteria for LB-MCI with participant with MCI without these features, healthy controls (HC), and participants with Parkinson's disease (PD) with and without MCI. Participants with MCI, HC, and PD were recruited as part of the COMPASS-ND study and underwent assessment of demographic and clinical features. Participants with MCI were classified as LB-MCI based on one or more criteria for LB disease. Plasma biomarkers were determined for amyloid species (Aß 42/40 ratio), tau-181, GFAP, and (NfL using Simoa. Groups were compared using ANOVA with post hoc comparisons. Age and sex adjustment was done in ANCOVA models. Among participants with MCI in COMPASS-ND, there were 159 (97 M/62 F) with MCI but without LB features and 105 (46 M/32 F) with one or more of the LB-MCI criteria. There were 161 HC (54 M/107 F), 79 PD (42 M/37 F), and 41 PD-MCI (35 M/7 F). The Aß 42/40 ratio and tau-181 differed between groups, with the LB-MCI group showing a lower Aß 42/40 ratio and higher tau-181 than the other groups. Post hoc comparison indicated that the ratio was lower in LB-MCI than HC, PD, and PD-MCI while tau-181 was higher than HC, PD, and MCI without LB features. GFAP and NfL did not differ across groups. Age and sex adjustment did not alter the main findings. In COMPASS-ND, participants with features of LB-MCI showed a biomarker profile suggestive of high Alzheimer co-pathology. Higher tau-181 suggests that these individuals might have a higher pathological burden than the other groups. The influence of other co-pathologies (i.e., vascular) and the influence on outcomes will be examined in this cohort. Future studies should examine for the presence of synuclein pathology across these groups.

Objective.Transcranial ultrasound stimulation (TUS) presents challenges in ultrasound wave transmission through the skull, affecting study outcomes due to aberration and attenuation. While planning strategies incorporating 3D computed tomography (CT) scans help mitigate these issues, they expose participants to radiation, which can raise ethical concerns. A solution involves generating skull masks from participants' anatomical magnetic resonance imaging (MRI). This study aims to compare ultrasound field predictions between CT-derived and MRI-derived skull masks in TUS planning.Approach.Five participants with a range of skull density ratios (SDRs: 0.31, 0.42, 0.55, 0.67, and 0.79) were selected, each having both CT and T1/T2-weighted MRI scans. Ultrasound simulations were performed using BabelBrain software with a single-element transducer (diameter = 50 mm,F# = 1) at 250, 500, and 750 kHz frequencies. CT scans were used to generate maps of the skull's acoustic properties. The MRI scans were processed using the Charm segmentation tool from the SimNIBS tool suite using default and custom settings adapted for better skull segmentation. Ultrasound was adjusted to target 30 mm below the skull's surface at 54 electroencephalogram (EEG) locations.Main Results.The custom setting in Charm significantly improved the Dice coefficient between MRI- and CT-derived masks when compared to the default setting (p< 0.001). The maximum pressure error significantly decreased in the custom setting compared to the default setting (p< 0.001). Additionally, the focus location error median across different SDRs averaged 2.32, 1.45, and 1.57 mm in default and 2.08, 1.38, and 1.44 mm in custom conditions for 250 kHz, 500 kHz, and 750 kHz respectively.Significance.MRI-derived skull masks offer satisfactory accuracy at many EEG sites, and using custom settings can further enhance this accuracy. However, significant errors at specific locations highlight the importance of carefully considering stimulation location when choosing between CT- and MRI-derived skull modeling.

BACKGROUND AND OBJECTIVES:Intensive speech therapy may improve recovery from poststroke aphasia. Further evidence suggests that pairing repetitive transcranial magnetic stimulation (rTMS) with intensive speech therapy might augment outcomes. This sham-controlled randomized clinical trial evaluated the efficacy of 1-Hz rTMS over the right pars triangularis combined with multimodality aphasia therapy (M-MAT) in chronic poststroke aphasia. METHODS:A parallel-group, double-blind, sham-controlled randomized clinical trial was conducted between April 2021 and May 2023 at an outpatient neurorehabilitation clinic. Individuals with chronic nonfluent aphasia after left middle cerebral artery stroke (>6 months from stroke) were enrolled and randomly assigned to receive either rTMS or sham stimulation combined with 35 hours of M-MAT over 10 days. The primary outcome was the Western Aphasia Battery aphasia quotient (WAB-AQ) measured at 3 weeks and 15 weeks. Intention-to-treat analysis examined treatment effects over time using linear mixed models. RESULTS:A total of 44 participants were randomized. Forty-three (mean [SD] age, 63.4 [12.3] years; 14 women [32.6%]) completed the intervention. Overall, WAB-AQ scores improved from baseline to 15 weeks regardless of rTMS allocation (mean difference 5.33, 95% CI 2.9-7.8, p < 0.001). We observed a significant group-by-time interaction (β = 0.31, p = 0.024), suggesting that those who received rTMS combined with M-MAT improved more over time than those who received sham. At 15 weeks, the rTMS group demonstrated significantly less word-finding difficulties and more complete and longer sentences with fewer pauses compared with sham as indicated by higher WAB-AQ scores (mean difference 4.1 points, 95% CI 0.6-7.6, p = 0.022). The change from baseline at 15 weeks was greater in the rTMS group (7.6 points, 95% CI 4.1-11.1) compared with sham (3.0 points, 95% CI -0.3 to 5.2; mean difference 4.6 points, 95% CI 0.6-8.6, p = 0.024). DISCUSSION:Intensive administration of M-MAT alone improves speech production in patients with chronic poststroke aphasia. Combining 1-Hz rTMS with M-MAT is associated with supplemental improvements in aphasia severity at follow-up. rTMS is a promising candidate as an adjuvant therapy to M-MAT. TRIAL REGISTRATION INFORMATION:ClinicalTrials.gov Identifier: NCT04102228. CLASSIFICATION OF EVIDENCE:This study provides Class III evidence that in patients with aphasia 6 or more months after a stroke, 1-Hz rTMS combined with intensive M-MAT improves WAB-AQ more than sham stimulation plus M-MAT.

Oxidative stress has been implicated in Parkinson disease (PD). Genes involved in PD, such as PRKN, PINK1, and PARK7, contribute to oxidative stress in dopaminergic neurons. The X-linked G6PD gene encodes glucose 6-phosphate dehydrogenase, an important regulator of oxidative stress. Recent studies suggested that alpha-synuclein aggregates may impair G6PD activity and contribute to dopaminergic neuron loss, and that G6PD mutations may independently increase the risk of PD. In this study, we aimed to examine the role of common and rare G6PD variants in PD across 6 cohorts, including 8,905 PD cases, 16,770 proxy cases, and 394,098 controls. These cohorts were analyzed after stratification by sex and then combined to account for the G6PD X-linked location. Using logistic regression, we did not identify significant associations for common variants in any of the cohorts. The optimized sequence Kernel association (SKAT-O) test was performed to assess the effect of rare variants (minor allele frequency <0.01) across six cohorts, followed by a meta-analysis using metaSKAT, also demonstrating lack of association. In conclusion, we did not find evidence for a role for G6PD in PD.

BACKGROUND:Mild behavioral impairment (MBI) is a syndrome characterized by the later-life onset of neuropsychiatric symptoms (NPS) and serves as a potential marker for dementia. In Parkinson's disease (PD), MBI has been associated with worse cognition, cortical atrophy, and altered connectivity. Unlike existing instruments that assess NPS in PD, the MBI Checklist (MBI-C) leverages sustained behavioral changes to identify patients at risk of cognitive impairment and neurodegeneration. The 34-item MBI-C has yet to be validated in PD. OBJECTIVE:This study assesses the MBI-C's psychometric properties in a multicenter Canadian PD sample and proposes a revised version optimized for PD. METHODS:A total of 406 PD patients from the Canadian Open Parkinson Network were assessed with the MBI-C to investigate its construct validity. Additional evaluations, including the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS), Neuropsychiatric Inventory (NPI), and Montreal Cognitive Assessment (MoCA), were implemented to examine the concurrent and criterion validities of the checklist. RESULTS:The original MBI-C exhibited considerable floor effects (24.9%). Exploratory factor analysis revealed a 5-factor 24-item MBI-C as consistent for PD (Cronbach's α = 0.856). All intrafactor correlations were statistically significant (P < 0.05), and convergent validity was found to be higher than divergent validity. Moreover, the MBI-C demonstrated moderate concurrent validity with the NPI (intraclass correlation coefficient [ICC]: 0.633, P < 0.001). The cutoff score associated with cognitive impairment on the revised instrument was 6|7. CONCLUSIONS:Compared to the original version, the revised MBI-C exhibited enhanced psychometric properties for measuring MBI in PD. It also demonstrated acceptable specificity when related to cognitive impairment. Future psychometric research should focus on capturing the subtlest manifestations of MBI in PD, examining patient-caregiver concordance, and addressing predictive validity for cognitive decline.

Purpose:Distributed learning is widely used to comply with data-sharing regulations and access diverse datasets for training machine learning (ML) models. The traveling model (TM) is a distributed learning approach that sequentially trains with data from one center at a time, which is especially advantageous when dealing with limited local datasets. However, a critical concern emerges when centers utilize different scanners for data acquisition, which could potentially lead models to exploit these differences as shortcuts. Although data harmonization can mitigate this issue, current methods typically rely on large or paired datasets, which can be impractical to obtain in distributed setups. Approach:We introduced HarmonyTM, a data harmonization method tailored for the TM. HarmonyTM effectively mitigates bias in the model's feature representation while retaining crucial disease-related information, all without requiring extensive datasets. Specifically, we employed adversarial training to "unlearn" bias from the features used in the model for classifying Parkinson's disease (PD). We evaluated HarmonyTM using multi-center three-dimensional (3D) neuroimaging datasets from 83 centers using 23 different scanners. Results:Our results show that HarmonyTM improved PD classification accuracy from 72% to 76% and reduced (unwanted) scanner classification accuracy from 53% to 30% in the TM setup. Conclusion:HarmonyTM is a method tailored for harmonizing 3D neuroimaging data within the TM approach, aiming to minimize shortcut learning in distributed setups. This prevents the disease classifier from leveraging scanner-specific details to classify patients with or without PD-a key aspect for deploying ML models for clinical applications.