Healthy Brains Healthy Lives (HBHL) Symposium 2024 | May 8-9th, 2024
Healthy Brains Healthy Lives (HBHL)
Published online: May 8-9th, 2024A DeepLabCut-derived framework for precise quantification behavioural dynamics in common marmoset
2Department of Neurology and Neurosurgery, Azrieli Centre for Autism Research, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
Corresponding Author: Justine Cléry, email: justine.clery@mcgill.ca
Abstract
In systems neuroscience, movement is a fundamental behavioral feature, which is indicative of the animal’s well-being and a potential disease status. To track animal behaviors, the usage of automatic machine-learning tools has become increasingly popular, such as the pose estimation program DeepLabCut (DLC). However, there is a limited toolkits to analyze the complex DLC behavioral data in an easy, accessible manner. Thus, we present a Python-based pipeline to facilitate complex behavioral analysis in freely moving marmosets based on the DLC-derived coordinates. In this study, marmosets were housed in family units within home cages, with their behaviors recorded via a Webcam-based system. With user-defined points of interest, we therefore investigated the recorded marmoset videos using the DLC to capture the motion and coordinates of different body parts. We analyzed the DLC-derived coordinates using the pipeline to measure continuous speed, inter-individual distance, and head posture consistency. Our results show that this pipeline enables quantitative predictions of movement patterns with high accuracy, applicable for both single and multiple marmosets’ projects. Depending on their activity levels and social relationships, we found that marmosets exhibit various speed intensity (e.g. resting, jumping), types of social interaction (e.g. play-fighting, friendship), and head posture. Overall, this approach enhances objectivity, reduces human errors, and minimizes repetitive works for behavioral analysis. Moreover, by comparing these results with a marmoset synucleinopathy model (Parkinson’s disease, dementia), this pipeline will allow us to characterize potential behavioral deficits to better understand the impact of synucleinopathy.
A genome-wide CRISPR/Cas9 screen identifies genes that regulate the cellular uptake of α-synuclein fibrils by modulating heparan sulfate proteoglycans
2Centre d’Excellence de Recherche sur les Maladies Orphelines – Fondation Courtois (CERMO-FC).
3The Neuro's Early Drug Discovery Unit (EDDU), Department of Neurology and Neurosurgery, Montreal Neurological Institute-Hospital, McGill University, Montreal, QC, Canada.
4McGill Parkinson Program, Department of Neurology and Neurosurgery, Montreal Neurological Institute and Hospital, McGill University, Montréal, QC, Canada.
5Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, ON, Canada.
6Division of Chemical Biology & Medicinal Chemistry, Eshelman School of Pharmacy, University of North Carolina, United States of America.
7Department of Chemistry and Chemical Biology, Rensselaer Polytechnic Institute, Troy, NY, Unites States of America.
8Department of Pharmacology & Therapeutics and Centre de Recherche en Biologie Structurale, McGill University, Montréal, QC, Canada.
Corresponding Author: Benoît Vanderperre, Edward A. Fon, email: vanderperre.benoit@uqam.ca; ted.fon@mcgill.ca
Abstract
Synucleinopathies are characterized by the accumulation and propagation of α-synuclein (α-syn) aggregates throughout the brain, leading to neuronal dysfunction and death. Understanding how these aggregates propagate from cell to cell in a prion-like fashion thus holds great therapeutic promises. Here, we focused on understanding the cellular processes involved in the entry and accumulation of pathological α-syn aggregates. We used an unbiased FACS-based genome-wide CRISPR/Cas9 knockout (KO) screening to identify genes that regulate the accumulation of α-syn preformed fibrils (PFFs) in cells. We identified key genes and pathways specifically implicated in α-syn PFFs intracellular accumulation, including heparan sulfate proteoglycans (HSPG) biosynthesis and Golgi trafficking. We show that all confirmed hits affect heparan sulfate (HS), a post-translational modification known to act as a receptor for proteinaceous aggregates including α-syn and tau. Intriguingly, KO of SLC39A9 and C3orf58 genes, encoding respectively a Golgi-localized exporter of Zn2+, and the Golgi-localized putative kinase DIPK2A, specifically impaired the uptake of α-syn PFFs uptake but not of tau oligomers, by preventing the binding of PFFs to the cell surface. Mass spectrometry-based analysis of HS chains indicated major defects in HS maturation in SLC39A9 and C3orf58 KO cells, explaining the cell surface binding deficit. Our findings establish these two genes as HSPG-modulating factors. Interestingly, C3orf58 KO human iPSC-derived microglia exhibited a strong reduction in their ability to internalize α-syn PFFs. Altogether, our data establish HSPGs as major receptors for α-syn PFFs binding on the cell surface and identifies new players in α-syn PFF binding and uptake.
Accuracy of FDG-PET scan to differentiate sporadic bvFTD from late onset psychiatric disorders
2Alzheimer Center Amsterdam, Neurology, Vrije Universiteit Amsterdam, Amsterdam UMC location VUmc, Amsterdam, The Netherlands
3Amsterdam Neuroscience, Neurodegeneration, Amsterdam, The Netherlands
4The University of Sydney, School of Psychology and Brain & Mind Centre, Sydney, Australia
5Technical University of Munich, School of Medicine and Health, Klinikum rechts der Isar, Department of Psychiatry and Psychotherapy, Munich, Germany.
6kbo-Inn-Salzach-Klinikum, Clinical Center for Psychiatry, Psychotherapy, Psychosomatic Medicine, Geriatrics and Neurology, Wasserburg/Inn, Germany.
7Department of Biomedical, Surgical and Dental Sciences. University of Milan, Milan, Italy
8Fondazione Ca’ Granda, IRCCS Ospedale Maggiore Policlinico, Milan, Italy
9The University of Sydney Brain and Mind Centre and Faculty of Medicine and Health School of Medical Sciences, Camperdown, NSW, Australia.
10The University of Sydney Brain and Mind Centre and Faculty of Medicine and Health School of Health Sciences, Camperdown, NSW, Australia. 11. McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
Corresponding Author: Ishana Rue, email: ishana.rue@mail.mcgill.ca
Abstract
State of the art: Positron emission tomography with [18F] fluorodeoxyglucose (FDG-PET) brain scan has shown strong efficacy to differentiate behavioral variant frontotemporal dementia (bvFTD) from other neurodegenerative disorders. However, the accuracy of FDG-PET to differentiate bvFTD from primary psychiatric disorders (PPD) remains understudied with prior studies reporting high false positive rates. Methodology: DIPPA-FTD has compiled a retrospective database with 508 sporadic bvFTD and 152 late-onset PPD cases, ages 45 and over, from five international sites. 226 cases had an FDG-PET scan completed at baseline with a visual rating of Yes, No or Ambiguous for accordance with a bvFTD diagnosis. The follow-up clinical diagnosis after at least 1 year was used as the gold standard to calculate the sensitivity and specificity of FDG-PET at baseline using two-way contingency tables. Results: There were 198 probable bvFTD cases (87.6%) and 28 PPD cases (12.4%) with a previous FDG-PET scan at follow-up. FDG-PET had a sensitivity of 87% and a specificity of 93% to identify bvFTD cases. The overall diagnostic accuracy was 88%. Within the PPD group, 7.1% were false positive scans and 14.3% percent had ambiguous scans. Conclusion: This study demonstrated an excellent diagnostic accuracy for FDG-PET for the differential diagnosis between sporadic bvFTD and late onset PPD in a multicentre retrospective cohort. The true false positive rate in PPD was low. These results support the use of FDG-PET as an important step in the diagnostic algorithm for ambiguous cases with late onset behavioural changes.
Advancing Neural Mass Models: Innovative Approaches for Conduction Delay Treatment and Enhanced Simulation Efficiency
2University of Electronic Science and Technology of China, Chengdu, Sichuan, China
Corresponding Author: Anisleidy Gonzalez-Mitjans, email: anisleidy.gonzalezmitjans@mcgill.ca
Abstract
Neural Mass Models (NMMs) are instrumental in exploring the intricate interactions among neuronal populations. However, existing NMMs, notably the classical Jansen and Rit (JR) model, encounter limitations, particularly concerning modularization and the treatment of delays. In this study, we address these shortcomings by introducing a novel approach to modeling realistic conduction delays, viewing delays as a spectrum of possible times rather than singular instances. A standout feature of our approach is the incorporation of disturbed conduction delays between neural masses via a Connectome Tensor (CT) formulation. Our proposed method aims to enhance the efficiency of NMM simulations through two key methodologies: the Local Linearization Method and strategic utilization of the Connectome Tensor. The suggested formulation is demonstrated by simulating with our Distributed delay NMM (DD-NMM) toolbox diverse configurations of JR cortical columns with different topologies and delays. The simulations scale from a single JR cortical column to a large-scale scenario with up to 1000 columns. The large-scale DD-NMM simulations reveal the network's high sensitivity to changes in the CT. Moreover, these simulations showcase a substantial reduction in computation time compared to conventional methods, rendering them approximately an order of magnitude faster and linearly scalable with the number of masses. We have developed algorithms and a public toolbox for seamlessly integrating high-dimensional NMMs with remarkable precision and computational efficiency. Through this work, we aspire to facilitate the integration of biophysical models for principled multimodal data fusion within the framework of the BigBrain Project.
Analysis of scRNAseq from Parkinson’s Disease Patients’ Human iPSC Derived Neurons and Microglia Grown in Pooled Cultures
2The Neuro, Early Drug Discovery Unit, Montreal Neurological Institute-Hospital, McGill University, Montreal, Quebec H3A 2B4, Canada
3Department of Human Genetics, Montreal Neurological Institute-Hospital, McGill University, Montreal, Quebec H3A 2B4, Canada
Corresponding Author: Edward A. Fon, Rhalena A. Thomas, email: ted.fon@mcgill.ca, rhalena.thomas@mcgill.ca
Abstract
Induced pluripotent stem cells (iPSC) enable the study of neurological disease pathways using single cell RNA sequencing (scRNAseq); yet, sequencing entire cohorts of patients remains prohibitively expensive, iPSC-derived brain cell cultures are laborious, and the reagents are costly, rendering cohort-scale analyses impractical. However, multiplexing, whereby cells from multiple subjects are pooled and differentiated in the same dish, could facilitate cohort scale experiments. Genetic demultiplexing could then assign pooled cells back to their subject-of-origin. However, the efficiency of demultiplexing algorithms and pooled culture conditions remain largely unexplored. For this pilot study, we used four iPSC lines, one healthy control and three individuals with PD, from the Parkinson’s Progressive Marker Initiative to generate Dopaminergic neurons and microglia. We compared scRNAseq from three culture conditions: 1) iPSC lines grown separately; 2) iPSC lines grown in one dish (iPSC pool); 3) iPSC lines differentiated into neural precursor cells or hematopoietic cells then pooled and grown in one dish (precursor pool). For genetic demultiplexing we developed Ensemblex, an accuracy-weighted, probabilistic ensemble genetic demultiplexing framework. We observed that subject representation was more balanced in precursor pools than in the iPSC pools. Correlation analysis and differential gene expression (DEG) analysis revealed that transcription is not significantly altered by pooling, indicating that this method could scale scRNAseq analyses. DEG analysis identified significantly altered pathways in the PD patient cultures compared to the control cultures. We provide tools and insights for cohort-scale multiplexing experiments, paving the way for personalized therapeutic interventions and precision medicine approaches.
Aperiodic Epileptogenic Dynamics in Pediatric Drug-Resistant Epilepsy
2Department of Pediatrics, Division of Pediatric Neurology, Montreal Children’s Hospital, McGill University, Montreal, QC, Canada;
3Department of Pediatric Surgery, Division of Neurosurgery, Montreal Children’s Hospital, McGill University, Montreal, QC, Canada
Corresponding Author: Eleanor Hill, email: eleanor.hill@mail.mcgill.ca
Abstract
Rationale: Despite the availability of anti-seizure medications, ~1/3 of epilepsy patients continue to suffer from seizures, and surgery remains the only potential curative treatment option. Such drug-resistant cases can extend our understanding of seizure generation and thus improve our identification and delineation of focal seizure onset zones (SOZ). The aperiodic components of neurophysiological activity are indicators of excitation/inhibition balance in brain circuits, which is disrupted during a seizure. We hypothesize that focal perturbations in aperiodic activity occur as a patient approaches a seizure and thus can act as a marker of the SOZ. Method: For 14 (6 male, 4-16 years) pediatric drug-resistant epilepsy patients who underwent presurgical evaluation, we applied the Spectral Parameterization Resolved in Time (SPRiNT) algorithm to extract dynamic aperiodic activity (offset and exponent) of seizures captured non-invasively with magnetoencephalography and intracranial electroencephalography (iEEG). SPRiNT was applied to each of the iEEG contacts and cortical MEG sources. Results: Preliminary results from iEEG recordings show that seizure onsets are preceded by a decrease of the exponent in regions involved in early ictal manifestations. These observations suggest a possible increase in excitatory activity immediately before seizure onset. Discussion: Dynamics in aperiodic activity have the potential to lead toward a new generation of temporal and spatial markers of seizure origination. Such markers would not require the labor-intensive review of recordings by clinical experts. Our approach may also provide a deeper mechanistic insight into the mesoscopic neurophysiological mechanisms of epilepsy and inform more specific and automatic markers of seizure generation.
Are severe cardiorespiratory events associated with hippocampal subfield development in very preterm-born infants?
2Integrated Program in Neuroscience, Faculty of Medicine, McGill University, Montreal, Canada
3Department of Neurology & Neurosurgery, Faculty of Medicine, McGill University, Montreal, QC, Canada
4Robarts Research Institute, University of Western Ontario, London, Canada
5MR Clinical Science, Philips Healthcare, Mississauga, Ontario, Canada
6Department of Medical Imaging, Division of Pediatric Radiology, Montreal Children’s Hospital, McGill University Health Centre, Montreal, QC, Canada
7Faculty of Education, Western University, London, Canada
8Western Institute for Neuroscience, Western University, London, Canada
9Pediatrics, Schulich School of Medicine & Dentistry, Western University, London, Canada
10Department of Pediatrics, Division of Neonatology, Montreal Children’s Hospital, Montreal, QC, Canada
11Department of Medicine, Division of Experimental Medicine, McGill University, Montreal, Qc, Canada
12School of Physical & Occupational Therapy, McGill University, Montreal, QC, Canada
Corresponding Author: Marie Brossard-Racine, email: marie.brossardracine@mcgill.ca
Abstract
Very preterm (VPT) infants have immature lungs and respiratory control centers at birth, which puts them at risk for cardiorespiratory events (CREs) and subsequent hypoxia during neonatal hospitalization. Studies have identified a negative relationship between frequent CREs and neurodevelopmental outcomes in VPT-born toddlers, but research examining their effect on brain development is lacking. Considering the vulnerability of the hippocampus to hypoxia, this study aims to examine the relationship between CREs and hippocampal subfield volumes in VPT infants at term-equivalent age (TEA). Infants born at <32 weeks of gestational age (GA) admitted to the NICU of the Montreal Children's Hospital were recruited. Participants completed a brain magnetic resonance imaging (MRI) at TEA. The cumulative number of severe CREs (apneas, bradycardias, and/or desaturations requiring intervention) from birth until 40 weeks GA were extracted from medical records. Hippocampal subfield volumes were obtained using ‘Hippunfold’, a validated automatic segmentation pipeline. Multiple linear regression analyses were performed to determine the effects of CREs on subfield volumes, controlling for GA at birth and at MRI. Data from 36 VPT infants were analyzed. A higher number of CREs was significantly associated with smaller bilateral CA1 volumes at TEA (left: p=0.03; right: p=0.008), as well as smaller right CA2/CA3 (p=0.02) and right total hippocampus volumes (p=0.02). Our findings suggest that VPT infants' hippocampal development is vulnerable to severe CREs during the neonatal period, particularly affecting the bilateral CA1 region. Upcoming analyses in a larger sample size will include additional clinical variables to better understand these complex associations.
Association between DBM-derived Atrophy Patterns and Cognition in Frontotemporal Dementia Variants
2Douglas Mental Health University Institute, Montreal, QC, Canada
3McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, QC, Canada
4Department of Psychiatry, McGill University, Montreal, QC, Canada
Corresponding Author: Amelie Metz, Mahsa Dadar, email: amelie.metz@mail.mcgill.ca, mahsa.dadar@mcgill.ca
Abstract
Frontotemporal Dementia (FTD) is a rare form of early-onset dementia. It encompasses diverse subtypes with distinct symptoms and brain atrophy patterns, including behavioral variant frontotemporal dementia as well as nonfluent variant and semantic variant primary progressive aphasias. Due to this heterogeneity, the diagnosis of FTD and its subtypes poses a diagnostic challenge to clinicians, leading to high rates of misdiagnosis. Here, we used deformation-based morphometry (DBM), a sensitive method for analyzing structural brain differences in Magnetic Resonance Imaging (MRI) scans, to further our understanding of the relationship between atrophy patterns and clinical features across FTD variants. We aimed to provide a robust and automated approach to diagnosis. To this end, we analyzed MRI data from 136 patients from the frontotemporal lobar degeneration neuroimaging initiative (FTLDNI) cohort. Using partial least squares (PLS), we correlated DBM-derived atrophy patterns with cognitive test scores, identifying significant latent variables explaining over 85% of covariance between cognitive performance and brain atrophy. These variables highlighted frontal, temporal, and subcortical brain networks associated with different cognitive domains, particularly language function. We were also able to discern group differences in the brain-behavior relationship among FTD subtypes. Utilizing these results, we applied machine learning to predict FTD subtypes. Our model achieved an accuracy of 88% with equally high sensitivity and specificity, outperforming previous attempts at automated FTD diagnosis. Our study underscores the strong link between DBM-assessed neurodegeneration and cognitive symptoms in FTD, suggesting the potential of DBM-based atrophy-cognition relationships as imaging biomarkers for assessing disease severity and subtyping FTD.
Associations Between Genetics of Parkinson's Disease, Brain Structure, and Behavioral Phenotypes
2Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec, Canada
3Douglas Mental Health University Institute, Montreal, Quebec, Canada
4Centre for Advanced Research in Sleep Medicine, Hôpital du Sacré-Cœur de Montréal, Montreal, Quebec, Canada
5Harvard Medical School, Cambridge, MA, USA
6Department of Human Genetics, McGill University, Montreal, Quebec, Canada
7Department of Psychiatry, McGill University, Montréal, Québec, Canada
Corresponding Author: Houman Azizi, email: houman.azizi@mail.mcgill.ca
Abstract
Background: Parkinson’s disease (PD) is associated with brain tissue loss detectable by Magnetic Resonance Imaging (MRI) (Chan 2007). Here we test the relationship between genetic risk for PD and MRI-derived fractional anisotropy (FA), cortical surface area (SA), and subcortical volume. We then study the relation between these neuroanatomical measures and behavioral phenotypes. Method: Demographics, behavioral, genomic and brain imaging data were obtained for 40,000 UK Biobank participants. The relationships between PD polygenic risk score and grey and white matter morphometry were assessed using linear regression. Partial least square (PLS) analysis was then used to investigate the behavioral phenotypes linked with brain features. Result: PD polygenic risk score was positively associated with cortical SA, subcortical volume, and white matter FA across the brain. The PLS analysis revealed education level, household income, and fluid intelligence as positively associated with these brain features, and multiple deprivation index as negatively associated. Conclusion: These results reveal a link between genetic susceptibility to PD and brain characteristics indicative of greater size of grey and white matter structures. This indicates that genes implicated in PD may also lead to increases in neuronal numbers and connections. These associations were in turn related to certain behavioral phenotypes. The findings are consistent with the view that an increase in neural density may make brains vulnerable to neurodegeneration.
Behavioural and Hippocampal Cell Type-Specific Gene Expression Profiles Associated With Maternal Immune Activation and Circadian Disruption
2Integrated Program in Neuroscience, McGill University, Montréal, Québec, H3A 2B4.
3Department of Psychiatry, McGill University, Montréal, Québec, H3A 1A1.
Corresponding Author: Danae Penichet, email: danae.penichet@mail.mcgill.ca
Abstract
Schizophrenia is a chronic, multifactorial mental disorder, for which prenatal exposure to maternal immune activation (MIA) triggered by an infection is a significant environmental risk factor. Because circadian disruption - such as chronic jet lag (CJL) - can affect schizophrenia by altering neuronal health, we aimed to characterize the behavioural phenotypes and cell-type specific molecular changes associated with exposure to MIA and CJL during adolescence. Two cohorts of pregnant dams were injected with either saline (control) or viral mimic poly(I:C) (MIA) on embryonic day 9.5. After weaning, all offspring were exposed to normal lighting conditions or CJL (6-hour phase advance every two days) for four weeks. One cohort was then subjected to behavioural testing, while the second cohort was sacrificed and their brains were collected for dorsal-hippocampal single-nucleus RNA sequencing (snRNAseq). While MIA decreased sociability(Crawley’s, p=0.0101) and open-field activity levels over time(p<0.001) and increased anxiety-like behaviour(elevated plus-maze, p=0.0036) and the acoustic startle response(Pre-pulse inhibition, p=0.0002) in males, females only showed a trending effect of MIA on the acoustic startle response. However, CJL significantly decreased activity over time(p=0.0080) and increased anxiety-like behaviour(p=0.0158) in females, but only slightly increased social novelty in males. We have identified the cell populations from the snRNAseq data of males and the differential gene expression analyses are underway to be presented at the conference. Our results show sexual dimorphisms whereby male and female mice display more schizophrenia-relevant behaviours after MIA and CJL exposure, respectively. Our snRNAseq data will reveal the cell-specific molecular pathways associated with these phenotypes.
Cannabis and Tobacco Co-Use and its Association with Striatal Brain Morphometry: Leveraging Data from the ENIGMA Addiction Working Group
2Douglas Research Center, Montreal, QC, Canada;
3Department of Psychiatry, University of Vermont, Burlington, VT, USA
4School of Behavioral and Brain Sciences, The University of Texas at Dallas, Richardson, TX, USA
5Cognitive Fitness Laboratory, Queensland Institute of Medical Research Berghofer Medical Research Institute, Herston, QLD, Australia;
6Neuroscience of Addiction and Mental Health Program, Healthy Brain and Mind Research Centre; School of Behavioural & Health Sciences, Faculty of Health Science, Australian Catholic University, Fitzroy, VIC, Australia;
7School of Psychology, University of Wollongong Australia, Wollongong, NSW, Australia
8School of Psychiatry, UMC Utrecht, Utrecht, CX, Netherlands;
9Neuroscience of Addiction Lab, Center for Substance use and Addiction Research, Department of Psychology, Education & Child Studies, Erasmus University, Rotterdam, PA, Netherlands;
10Departments of Psychiatry and Biobehavioral Sciences and Molecular and Medical Pharmacology, UCLA, Los Angeles, CA, USA;
11Department of Psychiatry, McGill University, Montreal, QC, Canada
Corresponding Author: Zac J. S. Yeap, email: je.yeap@mail.mcgill.ca
Abstract
Background: Cannabis and tobacco are frequently used addictive substances and daily co-use is common. Cannabis use is associated with greater striatal gray matter volume (GMV), while tobacco use is associated with lower striatal GMV. However, the effects associated with their combined use on striatal GMV remain unclear. Therefore, we investigated whether daily cannabis and tobacco co-use was associated with different patterns of striatal GMV compared to either substance alone or no substance use. Method: Pooling T1-weighted MRI scans from 10 ENIGMA Addiction sites yielded a sample of N=273. Four groups were examined: individuals with co-use (CT, n=45), cannabis-only use (CAN, n=28), tobacco-only use (TOB, n=60), and no use (controls, n=140). Cannabis groups used ≥0.5 joints/day and tobacco groups used ≥5 cigarettes/day. Using Freesurfer, GMV in the nucleus accumbens, caudate nucleus, and putamen were extracted. We employed 2x2 ANCOVAs controlling for age, sex, site, and alcoholic drinks/day. Since years of cannabis use differed between CT and CAN, and years of tobacco and daily tobacco use differed between CT and TO, these were controlled for in the analyses. False-discovery-rate correction was applied to control for multiple comparisons. Results: In the right nucleus accumbens, there was a significant interaction between cannabis and tobacco use (p=0.042). Main effects for cannabis (p=0.06) and tobacco (p=0.39) were not significant. Post-hoc comparisons revealed higher GMV in CAN than CT (p=0.045), TOB (p=0.028), and controls (p=0.001); no other group differences emerged. Conclusions: Among individuals with cannabis use, tobacco co-use may suppress cannabis-induced GMV increases in the nucleus accumbens.
Characterizing ARL10: Insights into Mitochondrial Dynamics, MDV Formation, and mTOR Signaling Pathways
Corresponding Author: Allison Keil, Jack Collier, Mai Nguyen, Heidi McBride, email: allison.keil@mail.mcgill.ca
Abstract
We recently unveiled a pyroptotic pathway where immune signaling leads to mtDNA exit from mitochondria within mitochondrial derived vesicles (MDVs) that target lysosomes. Inflammatory activation of pore-forming gasdermin proteins then assemble within mtDNA-containing lysosomes, leading to mtDNA leakage into cytosol, activation of DNA sensing cGAS/STING and pyroptotic cell death. A CRISPR-based screen mapped a series of Parkinsons related proteins acting at specific steps along this pathway, further implicating immune signaling pathways in the etiology of PD. To investigate the mechanisms regulating MDV release, I focused on ARL10, an uncharacterized mitochondrial outer-membrane anchored GTPase enriched within the MDV proteome. Initial silencing of ARL10 demonstrated an 80% reduction in steady-state MDVs carrying cargo for degradation in lysosomes. ARL10 knockout lines did not show a strong block in steady-state MDV transport, suggesting compensatory machineries could rescue the loss. Instead, preliminary experiments suggest a selective requirement for ARL10 in LPS driven mtDNA release in MDVs. Unexpectedly, ARL10 KO cells showed rapid cell growth with clear changes in the mTORC2/Rictor metabolic sensing pathways, underscoring a novel link between mitochondrial signaling, nutrient sensing and response to infection. To mechanistically dissect the function of ARL10 in these pathways, I developed biochemical approaches to map the GTP-binding effectors, GTP exchange factors (GEF) and GTPase activating proteins (GAP) that regulate ARL10 function at the mitochondrial outer membrane. Overall, my research aims to characterize ARL10’s functions in the context of MDV mechanisms, with the hope of shedding light on its potential impact in health and disease.
Characterizing Sex Differences in the Expression of Cannabis Withdrawal Symptoms During Extended Cannabis Abstinence
2Douglas Mental Health University Institute, Verdun, Quebec, Canada;
3Department of Psychiatry, McGill University, Montreal, QC, Canada
Corresponding Author: Gabriella Malamud, email: gabriella.malamud@mail.mcgill.ca
Abstract
Despite rising rates of cannabis use disorder (CUD), long-term abstinence treatments remain ineffective. Cannabis withdrawal may be an optimal treatment target because it is a strong predictor of relapse. Symptoms begin 24 hours after cessation, peak within one week, and dissipate following 28 days of abstinence. Limited research suggests that females experience more severe cannabis withdrawal compared to males. However, these studies were either cross-sectional or less than 28 days long. Since cannabis withdrawal can last for several weeks with fluctuations in severity, we aimed to determine if and when cannabis withdrawal severity differed between males and females during 28 days of abstinence. Males (n=9) and females (n=7) with CUD ages 18-55 with positive urine toxicology and no comorbid DSM-5 Axis 1 disorder were recruited. Participants underwent 28 days of cannabis abstinence supported by contingency management and weekly behavioural support sessions. Cannabis withdrawal symptoms were assessed at baseline (12-hours post-cessation) and then weekly using the Marijuana Withdrawal Checklist (MWC). Abstinence was verified using The Timeline Follow-Back, a self-report interview. Seven males and 7 females successfully abstained from cannabis for 28 days; 2 males relapsed. A repeated measures ANOVA revealed a significant main effect of time on MWC (F2.23, 26.79 =3.74, p = 0.03); the time x sex interaction was not significant (F2.23, 26.79 = 0.45, p = 0.66). Preliminary findings demonstrate that cannabis withdrawal follows the established trajectory for both males and females, with no significant sex differences in withdrawal severity present during 28 days of cannabis abstinence.
Cognitive performance and abnormal immune activity: Mediation analyses with brain morphology in the UK Biobank
2Montreal Neurological Institute, Montréal, Canada
Corresponding Author: Katie M. Lavigne, email: katie.lavigne@mcgill.ca
Abstract
Cognitive impairments and abnormal immune activity are both associated with various clinical disorders. C-Reactive protein (CRP) is a marker associated with inflammation yet its association with cognitive performance remains unclear. Moreover, mechanisms linking CRP to cognition are not yet established. Brain structure may well mediate this relationship: immune processes play crucial roles in shaping and maintaining brain structure, with brain structure and function driving cognition. To investigate these associations, we used the United Kingdom Biobank, a large cohort study with extensive assessments. With data from 39,200 participants, we aimed first to determine the relationship between CRP and cognitive performance, and second, to assess metrics of brain morphology as potential mediators in this relationship. After accounting for numerous potential covariates, we found CRP levels to have small, negative associations with fluid intelligence (b = − 0.03, 95%CI [-0.05,-0.02], pcor = .004), and numeric memory (b = − 0.03, 95%CI[-0.05,-0.01], pcor = .007). We found no evidence of brain morphology mediating these relationships (all |ab| < 0.001, all pcor > .55). These findings suggest that serum-assessed CRP is of marginal importance for cognitive performance in mid-to-late aged Caucasians. Our findings of a small but statistically significant association between CRP and cognitive performance provide context to previously inconsistent reports regarding this relationship. Moreover, the seeming lack of involvement of brain morphology suggests that other brain metrics (e.g., connectivity, functional activation) and/or other immune markers may be more pertinent to this relationship.
Decoding Structural Connectivity after Stroke for Clinical Use: Bypassing the Need for Diffusion Tensor Imaging.
2Lady Davis Institute for Medical Research, Montréal, QC, Canada;
3Canadian Platform for Trials in Non-Invasive Brain Stimulation, Montréal, QC, Canada;
4Department of Medical Psychology | Neuropsychology and Gender Studies, Center for Neuropsychological Diagnostics and Intervention, Medical Faculty of the University of Cologne, Cologne, Germany;
5School of Rehabilitation Sciences, University of Ottawa, Ottawa, ON, Canada;
6Music and Health Research Institute, University of Ottawa, Ottawa, ON, Canada;
7Institute of Neurosciences and Medicine INM-1, Research Centre Juelich, Juelich, Germany;
8Department of Family Medicine, McGill University, Montréal, QC, Canada;
9Department of Radiology, Weill Cornell Medicine, New York, NY, USA;
10Department of Computational Biology, Cornell University, Ithaca, NY, USA
Corresponding Author: Franziska E. Hildesheim, email: franziska.hildesheim@mail.mcgill.ca
Abstract
Background: Understanding connectome disruption is vital for predicting patient-specific functional recovery after stroke. Despite Diffusion Tensor Imaging (DTI) being the gold-standard for assessing structural connectivity, it is not routinely included in clinical imaging protocols. This study presents a model-based alternative approach for the evaluation of connectivity disruption, relying solely on routinely collected T1 images and bypassing the need for DTI data. The primary objective of this study is to demonstrate the clinical utility of this model-based approach by validating it against individual tractography results within a cohort of stroke patients. Methods: Individual tractography was performed using DTI data from 23 subacute aphasic stroke patients (mean age: 65.00±10.03 years, 11 females). Tractograms were overlaid with manually outlined dilated lesionmasks to calculate Change in Connectivity (ChaCo) scores as estimates of structural connectivity disruption in 13 language-relevant grey matter regions. Results were compared to ChaCo scores obtained with the model-based approach (Network Modification/NeMo tool), which involves overlaying lesionmasks on a tractogram reference set of 420 healthy subjects. Results: Linear regression analyses revealed significant positive correlations for connectivity disruption scores obtained through the model-based vs. individual tractography-based approach across all 13 brain regions (mean discrepancy: 8±2%). Angular gyrus and inferior frontal gyrus pars opercularis demonstrated the lowest ChaCo score discrepancies. Conclusions: The presented model-based approach provides a robust, DTI-independent alternative for assessing structural connectivity disruption, thereby enhancing accessibility and clinical feasibility of connectivity assessment in stroke patients. Integrating structural connectivity data into multivariate models holds promise for refining predictions of functional outcomes after stroke.
Disentangling the Effect of Brain Size from Sex Differentiated Aging Trajectories
2Douglas Mental Health University Institute, Montreal, QC, Canada
Corresponding Author: Aliza Brzezinski-Rittner, Mahsa Dadar, Yashar Zeighami, email: aliza.brzezinskirittner@mail.mcgill.ca, mahsa.dadar@mcgill.ca, yashar.zeighami@mcgill.ca
Abstract
Aging leads to loss of cortical and subcortical gray and white matter, and these changes differ between females and males (Bethlehem et al., 2022). Many volumetric sex differences in the brain are driven by differences in brain size independent of sex, and different methods have been developed to control for this effect (Sanchis-Segura et al., 2020). Here, we seek to compare these methods against a gold standard to evaluate their performance in disentangling the effects of sex from those derived by total intracranial volume (TIV) differences in regional aging trajectories. We modeled regional aging trajectories using FreeSurfer-based volumetric information based on the DKT atlas from the UK Biobank dataset, using raw values as well as values adjusted using three adjustment methods: the proportions method, the residuals method (Mathalon et al., 1993), and the power-corrected proportions (PCP) method (Liu, et al. 2014). We performed our analyses in four different subsamples, including an age and TIV matched gold standard sample (13,240 participants, equal number of males and females, matched within 0.2% difference in TIV and age), and one matched only by age. We examined how the model estimates change between these methods and matching strategies for different variables of interest. The estimated effect of sex diminished when modeling aging trajectories in the matched sample and the models using the residuals and PCP methods yielded similar estimates to those from the matched sample. Our gold standard approach provides a reference to assess the effect of the different adjustment methods.
E-Cog: An Online Training Platform for Psychological Interventions
2Department of Psychology, McGill University, Montreal, QC, Canada
3Department of Education and Pedagogy, Université du Québec à Montréal, Montreal, QC, Canada
4Department of Psychiatry, McGill University, Montreal, QC, Canada
5Douglas Mental Health University Institute, Montreal, QC, Canada
Corresponding Author: Martin Lepage, Ana Elisa Sousa, email: martin.lepage@mcgill.ca, anaelisa.fariasdesousa@douglas.mcgill.ca
Abstract
Online learning is a convenient and cost-effective solution to deliver training in healthcare. The increasing need for remote interventions in healthcare has emphasized the need for asynchronous remote training that maintains the quality of in-person training. This study outlines the design, development, and implementation of E-Cog, a novel online training platform supporting remote cognitive health intervention training. Following the ADDIE model, E-Cog comprises two certifications (cognitive remediation and metacognitive training) with online modules developed by mental health experts, either adapted from in-person training or created anew. The technological structure met predefined requirements, ensuring customization, optimal user experience, data security, accessibility, and gamification. Content and technical components underwent testing at different stages of development to enhance final delivery quality. Implementation of E-Cog is occurring within a multi-site trial to deliver remote cognitive interventions to individuals with psychosis across Canada. Five pilot users assessed the platform and certifications pre-release, consensually finding it easy to use, with trustworthy content, and expressing an average 86% likelihood to use it again/recommend it to a friend. Currently, 11 therapists have received training using E-Cog. Technical challenges related to content progression were identified and resolved, with ongoing improvements based on user feedback. E-Cog has been implemented as a tool to deliver remote training that is engaging, asynchronous, and cost-effective, maintaining the quality of in-person training. Successful uptake of mental health practitioners was observed during the first year of implementation. Next steps involve a qualitative assessment of the platform's usability and its impact on cognitive health intervention delivery.
Effects of Sustained Cannabis Abstinence on Depressive Symptoms in People with Cannabis Use Disorder: A Pilot Study
2Douglas Mental Health University Institute, Verdun, QC, Canada;
3Department of Psychiatry, McGill University, Montreal, QC, Canada
Corresponding Author: Lyne Baaj, email: lyne.baaj@mail.mcgill.ca
Abstract
Introduction: While people often use cannabis to cope with depression, longitudinal studies suggest that cannabis use is associated with the development of depressive symptoms. Therefore, it is important to determine if cannabis abstinence has a beneficial impact on depressive symptoms. Previous studies in adults with psychiatric comorbidities found that depressive symptoms improved with 28 days of cannabis abstinence. To study if these findings extend to adults without psychiatric comorbidities, we investigated the effects of 28 days of cannabis abstinence on depressive symptoms in adults with cannabis use disorder (CUD) and no psychiatric comorbidities. Methods: Adults (N=25; 18-50 years old) with CUD, a positive cannabis urine toxicology, and no co-occurring DSM-5 Axis 1 disorders were randomized using a 3:2 ratio to an abstinence group (n=16) or a cannabis-as-usual control group (n=9), respectively. Depression was assessed weekly with the Hamilton-Depression Rating Scale. Cannabis abstinence was encouraged using contingency management and behavioural support sessions and verified with the Timeline Followback self-report interview. Results: Fourteen participants in the abstinence group sustained 28 days of abstinence. A repeated-measures ANOVA revealed no change in depressive symptoms over time (F(4,84)=1.83, p=.15). However, a pattern emerged among abstinent participants demonstrating that depressive symptoms increased from baseline and peaked at 7-days post-abstinence before returning to baseline levels. Conclusion: Increased depressive symptoms at 7 days post-abstinence may be attributable to withdrawal effects. Given that symptoms returned to baseline levels, indicates that depression may not improve with 28 days of abstinence. Future studies should examine if depression improves with longer abstinence durations.
Empathy is associated with patterns of resting-state functional connectivity in presymptomatic genetic frontotemporal dementia: A GENFI study
2University College London, London, United Kingdom;
3Douglas Mental Health University Hospital, Verdun, QC
Corresponding Author: Shanny Foo, email: shanny.foo@mail.mcgill.ca
Abstract
BACKGROUND AND AIMS: A robust feature characterizing frontotemporal dementia (FTD) from other dementias are early deficits in social cognition. The salience and default networks, comprising groups of brain regions studied using resting-state functional connectivity (RSFC), are linked to the processing of emotionally salient stimuli and inferring others’ mental states, respectively. Recent studies have revealed differences across presymptomatic FTD groups in RSFC and empathy separately. The present study evaluates the relationship between RSFC and empathy in presymptomatic FTD. METHODS: The GENFI cohort recruited 840 presymptomatic adults (Mage=44y±13; 59%F, 41%M) including pathogenic mutation carriers C9orf72 (n=180), GRN (n=178), MAPT (n=72) and non-mutation carriers (NMC, n=410). RSFC data was processed using CONN and divided into networks using a parcellation method based on the Yeo 17 network solution. A subsample completed the modified Interpersonal Reactivity Index, a measure of empathy. Partial least squares analysis was conducted to evaluate group differences in patterns of RSFC and empathy-RSFC relationships. RESULTS: A distributed pattern of RSFC dissociated NMC from the mutation carrier groups, revealing a number of edge-level differences (20.08% covariance explained, p=.02). Notably, within-salience network RSFC contrasted C9orf72 from GRN (31.08% covariance explained, p=0.04). Empathy-related patterns of RSFC in both salience and default networks distinguished GRN from the other groups (23.95% covariance explained, p=.04). CONCLUSIONS: Early between group differences among presymptomatic mutation carriers allude to variability in disease progression and trajectory prior to onset of clinical FTD. Future longitudinal work examining the relationship between social cognition and RSFC change over time may provide greater sensitivity for identifying at-risk individuals.
Enhanced Cell Detection and Morphological Extraction for Cell Classification
2University of Zagreb, Zagreb, Zagreb,
3McGill Centre for Integrative Neuroscience (MCIN), Montreal,, QC,Canada
Corresponding Author: Alejandro Salinas-Medina, email: alejandro.salinas@mail.mcgill.ca
Abstract
We introduce an innovative approach for automating cell detection and extracting morphological features from one-micrometer resolution histological sections of the hippocampus, utilizing BigBrain data and a Partial Differential Equation model. The framework encompasses five stages. Each specifically designed to address challenges in cell analysis within histological staining data. The initial stage focuses on isolating a Region of Interest from images of the hippocampal area. An upsampling technique follows, employing a 3-degree spline interpolation to enhance image resolution by doubling the pixel density, thus preserving the integrity of the original data. During the third phase of the process, we employ a parallelized anisotropic diffusion process, which is grounded in the Perona-Malik model, across eight CPUs. This approach notably reduces the time required for diffusion and cell detection to less than one hour. The diffusion tasks effectively utilize CPU parallelization, while the cell detection is expedited using an NVIDIA A100 GPU. This novel configuration streamlines automation and will be part of an automated cell detection and classification pipeline, optimizing the combined capabilities of parallelized GPUs and CPUs. This optimization is crucial for fast processing of 7404 image slices. The fourth stage involves applying the watershed algorithm for precise cell segmentation within the densely packed hippocampal area. Finally, morphological features are extracted based on the segmentation output. This methodology marks a substantial leap forward in the field of automated cell detection and classification within high-resolution brain imaging. Its potential to significantly influence medical research and diagnostic practices is immense.
Event-related potential (ERP) differences between high and low schizotypal social anxiety in varied social contexts
2Douglas Research Center, Douglas Hospital, Montreal, QC, Canada.
3Integrated program in Neuroscience, Faculty of Health Sciences and Medicine, McGill University, Montreal, QC, Canada.
4Department of Psychiatry, Faculty of Health Sciences and Medicine, McGill University, Montreal, QC, Canada.
Corresponding Author: J. Bruno Debruille , email: bruno.debruille@mcgill.ca
Abstract
Schizotypy encompasses a range of traits in healthy individuals that resemble, albeit to a lesser degree, those observed in schizophrenia patients. Higher schizotypal personality (SPQ) scores are associated with eccentric characteristics and an increased susceptibility to schizophrenia compared to individuals with lower scores. EEG-event-related potential (ERP) studies have revealed differences, such as larger N400 ERP amplitudes in those with high SPQ-disorganization traits. However, it remains unclear whether these groups exhibit distinct ERPs in social settings, in the presence of strangers versus friends. Social anxiety (SA), a SPQ trait, may influence ERPs in social contexts at a preconscious level. To investigate this, a within-subject study was conducted with healthy participants (n = 30), who completed an image memorization task in three social contexts: alone, with a friend, and with a stranger. Their SA scores were assessed using the SPQ-BR questionnaire. Negative correlations were observed between SPQ-SA scores and N400 ERP measures in both friend (r = -0.51, p<0.01) and stranger (r = -0.50, p < 0.01) contexts. Such negative correlations were also obtained for N300 and LPP ERPs (r>-0.40, p<0.05). While no significant ERP differences were found across the social contexts, repeated-measure ANOVAs revealed that individuals with high SPQ-SA scores exhibited larger N300 and N400 amplitudes compared to those with low SPQ-SA scores in each context (p < 0.05). Conversely, LPP differences were observed in the opposite direction. These N400 and LPP differences could be interpreted in terms of inhibition processes and be linked to their socio-mental well-being.
Examining Differences in Multisensory Integration Between Cochlear Implantees and Normal Hearing Individuals Using Steady-State Evoked Potentials
2Centre for Interdisciplinary Research in Music Media and Technology (CIRMMT), Montréal, Canada ;
3McGill University, Faculty of Medicine and Health Sciences, Department of otolaryngology, head and neck surgery, Montréal, Canada ;
4Research Institute of the McGill University Health Centre (RI-MUHC), Montréal, Canada ;
5International Max Planck Research School on Neuroscience of Communication, Leipzig University, Leipzig, Germany ;
6Université catholique de Louvain, Louvain-la-Neuve, Belgique
Corresponding Author: Olivier Valentin, email: m.olivier.valentin@gmail.com
Abstract
Multisensory integration, the process of synthesizing information from various senses, is essential for complex cognitive tasks such as engaging in conversation amidst background noise or coordinating movements to music. Cochlear implants (CIs), neural prostheses that restore hearing in the profoundly deaf, have been behaviorally linked to altered multisensory integration. This alteration potentially hampers daily activities that require the integration of auditory, visual, and motor cues. To further elucidate multisensory integration in CI users, we investigated the temporal coordination of auditory, visual, and motor information in CI users and compared it with that of normal hearing (NH) controls. We recorded neural activity using high-density electroencephalography (EEG). Specifically, steady-state evoked potentials (SSEPs) were measured during two tasks: a passive perception task and an active finger-tapping task synchronized with four isochronous stimulus conditions—an auditory metronome, a visual metronome, both metronomes presented synchronously at the same tempo, and both presented asynchronously at different tempos. Our findings reveal significant SSEP discrepancies between CI users and NH controls, suggesting that CIs may alter the neural mechanisms of sensory-motor integration. Notably, SSEP topographies of CI users reveal the recruitment of auditory areas in processing visual information, while their responses to audiovisual stimuli indicate a predominant reliance on visual inputs over auditory inputs. These insights pave the way for future research to develop targeted rehabilitative strategies. Enhancing multisensory processing through tailored interventions could significantly improve the quality of life and functional independence of CI users.
Geometry of representations in artificial neural networks predicts visual learning specificity
2Mila - Quebec AI Institute
Corresponding Author: Amir Ozhan Dehghani, Shahab Bakhtiari, email: amirozhan.dehghani@mail.mcgill.ca, shahab.bakhtiari@umontreal.ca
Abstract
Visual perceptual learning (VPL) refers to the improvement of humans’ performance in visual tasks through practice. One of the hallmarks of VPL is specificity, which refers to failures of generalization to unseen test conditions. Humans show varying degrees of specificity among tasks, but the underlying mechanisms of this specificity and its variability are not well understood. We used a deep learning approach to gain an understanding of specificity in VPL. We proposed a theoretical framework, based on artificial neural networks (ANNs), for explaining and predicting VPL specificity. In particular, we hypothesized that visual tasks that are more aligned with the representational geometry of the model yield lower VPL specificity. We used ANNs pre-trained on object recognition, and quantified the representational geometry of the model using the Representational Dissimilarity Matrix. We simulated humans’ VPL by fine-tuning the ANNs on a set of visual tasks known as the same/different. In every task, a pair of images were presented to the model that were either the same or different. The model was trained to report if the images were the same or different. After training, we assessed specificity by moving images to new locations, and measured the performance of the model. Our preliminary results supported the main hypothesis: tasks that were more aligned with the representational geometry of the model led to lower learning specificity. Our results held across model architectures and objective functions. Interestingly, our hypothesis was more strongly supported in ANNs that were trained to be robust to noise perturbations.
Hippocampal alpha-synuclein seed shows regional vulnerability in MRI-derived brain atrophy patterns using a longitudinal mouse model of synucleinopathy
2Cerebral Imaging Center, Douglas Mental Health University Institute, McGill University, Verdun, Quebec, Canada
3Department of Biological & Biomedical Engineering, McGill University, Montreal, Quebec, Canada
4Department of Psychiatry, McGill University, Montreal, Quebec, Canada
5McGill Parkinson Program, Neurodegenerative Diseases Group, Department of Neurology and Neurosurgery, Montreal Neurological Institute-Hospital, McGill University, Montreal, Québec, Canada
6Early Drug Discovery Unit, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada
Corresponding Author: Janice Park, Stephanie Tullo, Daniel Gallino, Sara Tou, Medhinee M. Malvankar, Yohan Yee, Esther del cid-Pellitero, Wen Luo, Irina Shlaifer, Thomas M. Durcan, Edward A. Fon, M. Mallar Chakravarty, email: sung.park2@mail.mcgill.ca, stephanie.tullo@mail.mcgill.ca, dnlgallino@gmail.com, saroura157@gmail.com , medhinee123@gmail.com , yohan.yee@gmail.com , esther.delcidpellitero@mcgill.ca, wen.luo2@mcgill.ca, irina.shlaifer@mcgill.ca, thomas.durcan@mcgill.ca, ted.fon@mcgill.ca, mallar.chak@gmail.com
Abstract
Parkinson’s Disease (PD) and other synucleinopathies are characterized by aggregation of toxic alpha-synuclein (αSyn) within Lewy bodies. Extensive research suggests a prion-like spreading of αSyn, but the mechanism for neurotoxicity mediated by this spreading is unclear yet crucial to developing therapeutics that halt disease progression. Recent work from our group demonstrated magnetic resonance imaging (MRI)-derived signatures of widespread brain atrophy upon striatal inoculation of pre-formed αSyn fibrils (PFF) in the M83 transgenic mouse model expressing the human αSyn A53T mutation (Tullo et al., 2023), suggesting a prion-like spread of pathology. Here, we test the specificity of this spread by inoculating another highly connected network hub; the hippocampus. Hemizygous (hemi) M83 mice were injected with 2.5 μL saline (PBS) or PFF into the right dorsal hippocampus (n=~8 group/sex/timepoint). Mice underwent in vivo MRI scanning (T1-weighted images; 100 μm³ isotropic voxels; 7T Bruker Biospec) at -7, 30, 90, and 120 days post-injection. Neuroanatomical measures were obtained using deformation-based morphometry (DBM) with linear mixed effects model analysis at the voxel level to assess whole-brain volumetric change longitudinally. We observe a highly focal brain atrophy pattern in PFF-injected mice, notably in the bilateral hippocampi, as opposed to previously demonstrated widespread brain atrophy from traditional striatal inoculation studies. Our results demonstrate a limited degree of αSyn transmission from the hippocampus, suggesting that αSyn is particularly tuned to spreading through the motor network. Therefore, we propose that their transmission pathway is regionally selective and not solely attributable to structural connectivity.
Improving Quantification of Aperiodic (1/f) Dynamics: Bayesian Model Selection in SPRiNT
Corresponding Author: Benjamin Lévesque Kinder, email: benjamin.levesquekinder@mail.mcgill.ca
Abstract
Neural oscillations are central to the study of neurophysiology (Buzsáki, 2006). These oscillatory dynamics are composed of rhythmic and arrhythmic components. Although the rhythmic components of these signals have been studied extensively, the arrhythmic components have largely been disregarded as noise despite being behaviourally (Albouy et al., 2017) and pathologically (Molina et al., 2020) meaningful. SPRiNT (Spectral Parametrization Resolved in Time; Wilson et al., 2021) characterizes the rhythmic and arrhythmic dynamics of neural oscillations. This method allows users to quantify the evolution of the spectral parameters through time. But SPRiNT is prone to fitting far more spurious peaks than regular specparam. To remedy this shortcoming, we implemented Bayesian model selection into SPRiNT (ms-SPRiNT). We simulated 10'000 naturalistic time-series with both aperiodic and periodic dynamics. We found that ms-SPRiNT has a meaningfully higher peak positive predictive value with relatively small loss in peak sensitivity compared to SPRiNT. Independent of processing applied, the absolute error of the aperiodic exponent and aperiodic offset did not vary. As such, ms-SPRiNT marks the full maturity of SPRiNT as a tool that researchers in functional neuroimaging and clinical neuroscience can deploy in their analyses.
Is there evidence of cognitive decline 5+ years after the onset of psychosis? A systematic review and meta-analysis.
2Douglas Research Centre, Montreal, QC, Canada,
3Department of Psychiatry, McGill University, Montreal, QC, Canada,
4Copenhagen University Hospital, Copenhagen, Denmark
Corresponding Author: Joseph Ghanem, email: joseph.ghanem@mail.mcgill.ca
Abstract
Psychotic disorders are characterized by marked cognitive deficits that are apparent prior to illness onset and persistent thereafter. However, whether these cognitive impairments decline, stabilize, or improve in the long-term remains contentious. Previous systematic reviews and meta-analyses were focused on the first 5 years following psychosis onset and identified improvements in cognition that appeared to reflect practice effects. Recent studies with follow-ups of a decade or more have suggested that cognitive decline may be apparent in some cognitive domains, but others observed that cognition stabilized over time. The present review aimed to evaluate the progression of cognition over time in studies with follow-ups of 5+ years. Twenty studies comprising 2412 participants and evaluating 7 cognitive domains were included in the present study. Global cognition remained stable over time (g = .06, 95% CI -.05 -.18), and so did the individual cognitive domains of verbal memory (g = .07, 95% CI -.08 - .23), visual memory (-.08, 95% CI -.31 -.15), working memory (g = .03, 95% CI -.12 - .18), speed of processing (g = .19, 95% CI -.10 -.47), reasoning and problem-solving (g = .13, 95% CI -.13 - .38), and verbal fluency (g = .09, 95% CI -.04 - .21). These findings indicate that neurocognition is stable after the onset of psychotic disorders and does not decline over time, lending support to the neurodevelopmental model of psychotic disorders.
Longitudinal Inference of Multimodal Cortical and Hippocampal Network Connectivity in Psychotic Disorders
2Douglas Research Centre, Montreal, QC, Canada;
3Department of Biological and Biomedical Engineering, McGill University, Montreal, QC, Canada;
4Department of Computer Science, University College London, London, United Kingdom
Corresponding Author: Katie M. Lavigne, email: katie.lavigne@mcgill.ca
Abstract
Novel machine-learning evidence in psychosis suggests two distinct disease progression patterns of volume atrophy, with one pattern commencing in the hippocampus. It remains to be explored whether this extends to the disease progression of multimodal hippocampal connectivity. Morphometric and functional MRI data was sampled from 175 and 66 patients with first episode (FEP) and 117 and 51 controls respectively. For both modalities, we correlated hippocampal and cortical measures, and then grouped the regions into eight modules (one hippocampal module and seven Yeo networks). Intermodular connectivity was derived through the graph-theoretical participation coefficient, and the average participation coefficients of all eight modules were used as input for two separate modality-specific analyses with SuStaIn. SuStaIn is a machine-learning algorithm which merges disease progression modeling and clustering, uniquely suited to identify connectivity progression patterns across psychosis subtypes. Following 10-fold cross-validation, SuStaIn identified morphometric and functional models with three subtypes each. In both models, Subtype 0 included individuals with normal-range connectivity on all markers. Across modalities, Subtype 1 progressed from decreased cortical connectivity towards decreased hippocampal connectivity and then increased cortical connectivity in other networks, while Subtype 2 started with increased connectivity in cortical networks, followed by reduced hippocampal connectivity and further deteriorations in cortical network connectivity. To conclude, hippocampal connectivity was the first to decrease after an increase in cortical connectivity across modalities. These findings replicate and extend prior work on volume atrophy toward multimodal connectivity progressions. Future work should investigate the relationship between these subtypes and clinical features of psychosis.
Multimodal Precision Neuroimaging of the Individual Human Brain at Ultra-High Field
2McConnell Brain Imaging Centre, Montreal Neurological Institute and Hospital, Montreal, QC, Canada
3Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Germany
4University of Oxford, Oxford, UK
5Institute for Neuroscience and Medicine (INM-7), Forschungszentrum Juelich, Juelich, Germany
6Queens University, Kingston, ON, Canada
Corresponding Author: Donna Gift Cabalo, email: donna.cabalo@mail.mcgill.ca
Abstract
Neuroimaging has advanced our understanding of the human brain by allowing non-invasive examination of brain structure and function. Nevertheless, human MRI research has predominantly centred around group-averaged data, which limits the specificity and clinical utility that MRI can offer[1,2]. Precision neuroimaging facilitates individualized mapping of brain structure and function through the use of repeated and prolonged scans[1,2]. Harnessing ultra-high field (UHF) neuroimaging at magnetic field strengths of 7Tesla, can further enhance spatial and temporal resolution. Here, we provide a multimodal precision neuroimaging dataset that capitalizes on multiple sessions of 7T MRI. Our imaging protocol was implemented at the MNI and data were acquired on a 7T Terra scanner. Ten healthy subjects (4M/6F, age=26.8±4.61) underwent three imaging sessions, each consisting of distinct structural and fMRI protocols. Our UHF-MRI data were processed with micapipe_v0.2.3[3], a surface-based processing software. Structural scans included: (i) T1w, (ii) T1 relaxometry (T1), (iii) DWI, (iv) magnetization transfer and (v) T2*-weighted multi-echo gradient echo. Multi-echo fMRI scan included (i) rs-fMRI, (ii,iii,iv) multi-state task-based fMRI with episodic encoding/retrieval and semantic tasks and (v) movies. In addition to the anonymized raw data which is now available at https://osf.io/mhq3f/, we will release fully processed data for each modality. Our open-access precision UHF dataset promises to become a key resource for researchers aiming to advance our understanding of structure-function relationships in individual human brains and is instrumental in the development of novel image processing and analysis methodology. [1]Gordon (2017). Neuron, 95(4) [2]Poldrack (2017). Neuron, 95(4), 727–729 [3]Cruces, (2022). NeuroImage, 263, 119612
Neural correlates of invariant coding for naturalistic stimuli in the electrosensory system
Corresponding Author: Amin Akhshi, Anmar Khadra, Maurice Chacron, email: amin.akhshi@mail.mcgill.ca, anmar.khadra@mcgill.ca, maurice.chacron@mcgill.ca
Abstract
Understanding how neurons process sensory information for successful environmental interaction remains a central problem in neuroscience. In particular, animals need to recognize various representations of sensory inputs from the same "object" under different conditions. This is thought to be achieved in the brain by having neurons respond in a similar manner (i.e., invariantly) to sensory input through identity-preserving transformations. While such invariant representations have been observed across systems and species, the mechanisms underlying this phenomena remain poorly understood. In our study, we explored how burst firing contributes to the invariant representation of electrocommunication signals ("chirps") in the electrosensory system of weakly electric fish. We used Neuropixels probes for multi-unit recordings from ELL pyramidal cells. Our findings reveal that, at the population level, burst firing leads to more reliable and similar responses across different chirps, resulting in a more invariant representation than that obtained by considering the entire spiking activity. Additionally, we developed a biophysical model of ELL pyramidal cells to validate the intrinsic mechanisms leading to our findings, highlighting the crucial role of somato-dendritic interactions in generating invariant representations. Finally, we assessed how downstream torus neurons in the hierarchy of the electrosensory circuit decode these invariant representations by developing and training a deep neural network (DNN) optimized for invariance. Overall, our results show a novel function for burst firing in establishing invariant representations by ELL pyramidal cell populations of natural electrosensory communication stimuli and suggest that such representations are decoded by downstream neurons to further optimize invariance.
Neuromelanin-sensitive MRI in cannabis use disorder and first episode schizophrenia.
2Douglas Research Institute, Verdun, Quebec, Canada.
3Robarts Research Institute & Lawson Health Research Institute, London, Ontario, Canada.
4Department of Medical Biophysics, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada.
5Department of Psychiatry, Dalhousie University, Halifax, Nova Scotia, Canada.
6Renaissance School of Medicine at Stony Brook University, Stony Brook, New York, USA.
Corresponding Author: Jessica Ahrens, email: jessica.ahrens@mail.mcgill.ca
Abstract
Adolescent cannabis use has been associated with increased risk of schizophrenia. Although the mechanism linking cannabis use disorder with schizophrenia remains elusive, aberrations in dopamine turnover is suspected to play a role, but findings to date have been contradictory. We investigated this relationship using neuromelanin-sensitive MRI (neuromelanin-MRI), a putative proxy measure of dopamine system function that has previously shown alterations in schizophrenia and substance use disorders. Twenty-five participants with cannabis use disorder (CUD) and 36 participants without CUD (nCUD) participated in the study. 28 of these participants had been diagnosed with first episode schizophrenia (FES). We collected a neuromelanin-MRI scan from the substantia nigra (SN) and tested the association of CUD and FES diagnoses with neuromelanin-MRI signal using robust linear regression analysis. CUD was associated with elevated neuromelanin-MRI signal in a cluster of ventral SN voxels (353 of 2060 SN voxels, pcorrected=0.023, permutation test). FES diagnosis was not associated with a significant alteration in SN NM-MRI signal. Within a SN subregion previously shown to be associated with severity of untreated schizophrenia, CUD participants showed elevated signal compared to nCUD participants (t56=2.01, p=0.049). This indicates an increased dopamine turnover in subregions relevant to schizophrenia risk in individuals with cannabis use disorder. Cannabis-associated elevation of dopamine turnover may contribute to the risk of schizophrenia in long-term users of cannabis. The relationship between neuromelanin-MRI signal in the SN and dopamine turnover in other brain regions critical for schizophrenia requires further clarification.
Neurotransmitter signaling networks constrain disease-specific patterns of cortical vulnerability
Corresponding Author: Filip Milisav, email: filip.milisav@mail.mcgill.ca
Abstract
The development of biologically resolved wiring diagrams is fundamental to the study of brain function and dysfunction. Several studies indicate that the brain’s structural network architecture mediates the propagation and the expression of brain disease. However, the susceptibility of brain regions to disease is also believed to be shaped by their physiochemical makeup. Here, we develop a framework for annotating structural network edges using chemoarchitectural features of 7 classical neurotransmitter systems to build a signaling-informed connectome. Neurotransmitters are the fundamental molecular units of neural signaling, and together with neurotransmitter receptors, they modulate neuronal excitability and mediate the propagation of electrical impulses. To account for the spatial patterning of neurotransmitter synthesis and neurotransmitter receptors, and relate it to inter-regional anatomical connections, we consider a whole-brain atlas of 17 receptors acquired from PET tracer images and gene expression data from the Allen Human Brain Atlas for rate-limiting enzymes associated with 7 corresponding neurotransmitters (acetylcholine, glutamate, GABA, dopamine, norepinephrine, serotonin, and histamine). Specifically, we weight each edge of the structural brain network by the product of gene expression and receptor density in the corresponding sender and target regions, respectively. The resulting neurotransmitter signaling networks provide a transmitter-receptor resolved perspective on inter-regional synaptic signaling. Compared to structure alone, they improve prediction of inter-regional functional connectivity and patterns of abnormal cortical thickness across 8 psychiatric disorders. Altogether, our results indicate that annotating brain structure with neurotransmitter chemoarchitecture can help in building more veridical representations of neural signaling, with potential for disease modelling.
NeuroWave Alert: Harnessing EEG Signals for Real-time Fatigue Detection in Drivers
2Department of Biomedical Engineering, University of Montreal, Montreal, QC, Canada
Corresponding Author: Sujata Sinha, Anuj Saini, email: sujata.sinha@mail.mcgill.ca, anuj.saini@umontreal.ca
Abstract
Drowsy driving, a significant cause of road accidents, often stems from driver fatigue. A promising approach to tackle this issue involves using EEG signals, which reflect an individual's sleep and wakefulness states. In this study, we developed a machine learning (ML) model to detect fatigue based on open-source EEG data obtained from the NeuroSky MindWave headset (Mohamed, 2023). The data, in microvolts squared per hertz (μV²/Hz), were captured from a single frontal-lobe electrode while 3735 driver-participants were in sleepy and awake conditions. After preprocessing, i.e., data scaling using StandardScaler and an 80:20 train-test split, ML models, such as Logistic Regression, Decision Tree, Bagging Classifier, Random Forest, and LSTM were evaluated. While most of these showed signs of overfitting, the LSTM outperformed others, with 81% accuracy on training data and 76% on the test. This model consisted of an initial LSTM layer with 128 units, a dropout layer for regularization, and a final sigmoid activation layer for binary classification. It was optimized using the Adam optimizer and binary cross-entropy loss over 100 epochs to strike a balance between learning efficiency and generalizability. The results showed that the top three contributing features to the model were highGamma (18.3%), delta (14.1%), and highBeta (11.9%), indicative of complete wakefulness. Surprisingly, attention (6%) contributed the least, suggesting that awake drivers may experience attention lapses due to fatigue or even fall asleep while driving. These findings thus strengthen the potential of EEG for detecting not only fatigue in drivers but also real-time mental states in individuals.
Nipoppy: a framework for the organization and decentralized processing of neuroimaging-clinical data
Corresponding Author: Michelle Wang, Jean-Baptiste Poline, email: michelle.wang6@mail.mcgill.ca, jean-baptiste.poline@mcgill.ca
Abstract
The organization and processing of neuroimaging-clinical datasets can be time-consuming and error-prone. Lack of methodological detail and provenance recording can also hinder reproducibility efforts. We introduce Nipoppy, a flexible and open framework that can help achieve decentralized, reproducible data curation and processing for multimodal datasets. The framework consists of 1) a data organization specification for raw and processed, imaging and non-imaging data, and 2) a standardized workflow process starting from raw imaging data and ending with imaging-derived phenotypes (IDPs). We provide a user-friendly software package to work within the framework, with tools to convert raw scanner output to the Brain Imaging Data Structure (BIDS) standard, process neuroimaging data with existing or custom pipelines, track processing completion status, and extract IDPs from pipeline outputs. We have also developed an interactive web dashboard for visualizing data availability and processing statuses. Nipoppy has successfully been used to process longitudinal Parkinson’s disease cohorts from the Parkinson’s Progression Markers Initiative, the Quebec Parkinson Network, and the National Institute of Mental Health and Neurosciences (India). All three datasets have been processed with pipelines such as fMRIPrep, MRIQC and TractoFlow. Nipoppy facilitates consistent organization and processing of large studies, which can help simplify data sharing between so-called “data silos”. Nipoppy also creates files compatible with the Neurobagel ecosystem for distributed dataset harmonization and search (https://neurobagel.org/), which would enable easy decentralized discovery and sharing of IDPs in the future.
POLR3-related Leukodystrophy: A Qualitative Study on Parents’ Experiences with the Healthcare System
2Department of Neurology and Neurosurgery, McGill University, Montréal, QC, Canada;
3Université de Montréal, Montréal, QC, Canada;
4Adaptation, Resilience, and Change Research Unit, Université de Liège, Liège, Belgium;
5Department of Pediatrics, McGill University, Montréal, QC, Canada;
6Department of Human Genetics, McGill University, Montréal, QC, Canada;
7Division of Medical Genetics, Department of Specialized Medicine, McGill University Health Centre, Montréal, QC, Canada
Corresponding Author: Geneviève Bernard, email: genevieve.bernard@mcgill.ca
Abstract
RNA polymerase III-related hypomyelinating leukodystrophy (POLR3-HLD) is a rare, genetically-determined, neurodegenerative disorder affecting the white matter of the central nervous system. It is caused by biallelic pathogenic variants in genes encoding subunits of RNA polymerase III; POLR3A, POLR3B, POLR1C, POLR3K, and POLR3D. POLR3-HLD is known colloquially as 4H leukodystrophy due to its hallmark clinical features of hypomyelination, hypodontia, and hypogonadotropic hypogonadism. Disease onset typically occurs in childhood and runs a progressive, debilitating, and often fatal course. Patients with this disorder require complex and specialized care, however, due to its rarity and limited widespread awareness, parents are often required to assume additional roles as experts and advocates for their child(ren). Here, we aimed to understand parents’ experience navigating the healthcare landscape and identify potential targets for improvement. 19 semi-structured interviews were conducted with an international cohort of 24 parents to obtain broad perspectives and were analyzed using reflexive thematic analysis to identify patterns of themes that address the research question. Four themes were identified: existing barriers in accessing care, limited knowledge in diagnosis and care, parents as experts and advocates of their child(ren)’s care, and perceived superior care by leukodystrophy specialists. Many parents expressed feeling alone and uncertain, with little guidance provided to them. They also identified perceived gaps in care and challenges they faced but found comfort when treated by leukodystrophy experts in specialty clinics. This study will help better inform healthcare providers, administrators, and policymakers to expand and improve access to quality care for POLR3-HLD patients and their families.
Reading efficiency under binocular and monocular reading conditions
2Integrative Program in Neuroscience, McGill University, Montreal, QC, Canada.
3Department of Ophthalmology, McGill University, Montreal, QC, Canada.
4Research Institute at the McGill University Health Centre, Montreal, QC, Canada.
Corresponding Author: Dasha Vanichkina, Alexandre Reynaud, email: darya.vanichkina@mail.mcgill.ca, alexandre.reynaud@mcgill.ca
Abstract
In the past, it has been shown that normally sighted individuals do not significantly benefit when reading with two eyes versus one. However, this question has been mainly evaluated through eye-tracking methods. Despite the lack of differences previously found, we believe that there may be different patterns of brain activity during binocular and monocular reading. Thus, to evaluate this theory, we applied a novel random temporal sampling method which allows us to make inferences about brain oscillation patterns. In this study, we investigated whether there were differences in reading efficiency along the temporal sampling domain during binocular and monocular reading. To quantify participants’ reading efficiency, we used the random temporal sampling technique at an accuracy level of 50%. On each trial, participants read the target word to the best of their ability, as varying levels of temporal noise impeded their reading, and they attempted to sort the word into one correct category out of four possibilities. If participants were reading monocularly, one of their eyes was covered with an eyepatch. We saw that participants generally found the monocular condition to be harder, which was reflected by their superior binocular performance. Therefore, we can conclude that there indeed may be differences in brain oscillations during binocular and monocular reading. Previously, the random temporal sampling method has been used to detect differences in the brains of individuals diagnosed with ADHD; hence, we believe that it can also be used to identify biomarkers of other neurological disorders such as amblyopia.
Reproducible cleaning and transformation of a historical research-clinical dataset: methods developed, and lessons learned by the Douglas Neuroinformatics Team
2Integrated Program in Neuroscience, McGill University, Montreal, QC, Canada
3Department of Biology and Biomedical Engineering, McGill University, Montreal, QC, Canada
4Department of Psychiatry, McGill University, Montreal, QC, Canada
Corresponding Author: Weijie Tan, Gabriel Devenyi, email: weijie.tan.comtl@ssss.gouv.qc.ca, gabriel.devenyi@douglas.mcgill.ca
Abstract
Measurement-Based Care generates data to improve treatment outcomes and facilitate research. However, manual data entry into proprietary software like SPSS poses accessibility challenges for collaborating researchers lacking licenses and introduces errors and inconsistencies due to weak data models. Manually extracting data for sub-cohort studies introduces further potential for inaccuracies, including inadvertent data inclusion or exclusion. Finally, related studies implemented over many years under such conditions can result in study design drift, requiring post-hoc harmonization. Here we present the results, and lessons learned from the cleaning and integration of a retrospective 15-year research-clinical dataset for public databanking. We developed a methodology for converting and cleaning a historical dataset of 88 SPSS files into an semi-long dataframe with time-based longitudinal representation. The resulting data table features clear variable names along with a metadata dictionary containing variable types, descriptions, value labels, and expressions for calculated variables. The development process addressed challenges due to inconsistent variable naming, insufficient dataset documentation, missing data, and data conflicts. From this experience, we present several valuable lessons: the importance of imposing data structure from the start, producing meticulous documentation, preserving digital copies of paper tests, implementing effective onboarding processes for new team members managing data collection, and comprehensive date logging for all events. These insights inform future development, enhancing efficiency, transparency, and resilience in managing datasets. Our streamlined method supports research analysis through automated script execution, overcoming software licensing restrictions and seamlessly integrating into ongoing data collection efforts to enhance data integrity and facilitate longitudinal research endeavors.
Rescue of Severe Synaptic Deficits in Human iPSC-Derived Neurons Across the Mucopolysaccharidosis III Spectrum with a Small Molecule Drug, AVP6.
2Department of Anatomy and Cell Biology, McGill University, Montreal, H3A 0C7, QC, Canada; 3 Phoenix Nest Inc, Brooklyn, NY, 11232, USA. 4 Montreal Neurological Institute and Hospital, McGill University, Montreal, H3A 0C7, QC, Canada
Corresponding Author: Travis Moore, Patricia Dubot, Poulomi Bose, Jill Wood, Thomas Durcan, Alexey Pshezhetsky, email: patriciadubot@gmail.com, poulomeebose@gmail.com, jwood@phoenixnestbiotech.com, poulomeebose@gmail.com, alexei.pchejetski@umontreal.ca
Abstract
Sanfilippo Disease, or Mucopolysaccharidosis (MPS) III, is a spectrum (A-D) of rare inherited neurological lysosomal diseases caused by deficiencies in one of four enzymes that catabolize heparan sulphate. Sanfilippo causes a severe neurocognitive decline in affected children around 2-3 years old that leads to premature death. Currently, MPS III lacks any effective neurological treatments. Previously, our research identified several neuron-specific biomarkers of pathogenesis in the MPS IIIC mouse models for the excitatory and inhibitory neuronal systems and decreased brain-derived neurotrophic factor (BDNF). However, these deficits have yet to be confirmed in a human-relevant or patient-specific neuronal model of Sanfilippo disease. We have also identified a potential small molecular therapeutic, designated “AVP6,” with both neuroprotective and neurotrophic effects by increasing BDNF. The project aims to develop in vitro human-relevant neuronal models of induced pluripotent stem cells (iPSCs) of MPS III patients (A, B, and C) and healthy controls (age/sex-matched) to characterize the disease and test potential therapeutics. Utilizing iPSCs from MPS III patients, the cells have been differentiated into cortical neurons and characterized. The MPS III neurons were shown to retain their disease-specific primary enzymatic deficits, showed an increased LAMP2+ area, suggestive of lysosomal storage, and revealed significant defects of protein markers for glutamatergic (VGLUT1/PSD-95) and GABA-ergic (VGAT/Gephyrin) synapses and BDNF. Treatments with AVP6 during maturation significantly increased or restored levels of all synaptic proteins and BDNF. Our results demonstrate the importance of iPSC-derived neurons, their ability to characterize severe synaptic deficits and evaluate AVP6s as a potential therapeutic.
Sex-Specific causal dynamic between Insulin resistance and MDD, a bidirectional Mendelian randomization study
2Douglas Mental Health University Institute, McGill University, Montreal, QC, Canada;
3Ludmer Centre for Neuroinformatics and Mental Health, Douglas Research Centre, McGill University, Montreal, QC, Canada;
4Department of Psychiatry, Faculty of Medicine, McGill University, Montreal, QC, Canada
Corresponding Author: Patricia P.Silveira, email: patricia.silveira@mcgill.ca
Abstract
Aims: Evidence from various studies has shown a bidirectional association between major depressive disorder (MDD), Insulin resistance (IR), and related diseases, which varies between sexes and ancestries. There has also been evidence suggesting that IR influences response to antidepressant treatment. We conducted a sex-specific two-sample bidirectional Mendelian randomization (MR) study to assess the causal associations of MDD with Insulin resistance measured through the TG: HDL-C ratio and vice versa. Furthermore, we also performed another two-sample Mendelian analysis to assess the causal influence of IR on anti-depressant response. Methods: We obtained summary-level statistics for MDD, antidepressant response, and insulin resistance from corresponding published large genome-wide association studies of individuals mainly of European ancestry and partially replicated the analyses using available summary data from studies of individuals of East Asian descent. The random-effects inverse-variance weighted method was used for the main analyses. Results: Genetic liability to MDD was significantly associated with Insulin resistance both generally and sex-specifically, while the causal effect of Insulin resistance on MDD is only significant in females. We found limited evidence supporting the causal effects of insulin resistance on antidepressant response. None of the previous associations were found using summary data with East Asian subjects. Conclusions: The present study consolidated MDD as a potential risk factor for insulin resistance and that insulin resistance plays a sex—and ancestry-specific role in MDD pathology. Together, these findings could contribute to further our understanding of the comorbidity between MDD and IR-related diseases, allowing for more individualized treatment and diagnosis.
Studying light-evoked retinal responses following optogenetic vision therapy
Corresponding Author: Stuart Trenholm, email: stuart.trenholm@mcgill.ca
Abstract
Background: Retinal degenerative diseases are a leading cause of vision loss and arise from the loss of photoreceptors. Following degeneration, other cells in the retina remain intact and could be targeted with optogenetics as a therapeutic strategy. The healthy retina sends out many channels of visual information, via functional types of retinal ganglion cells (RGCs), coding different visual features, such as motion, contrast, etc. We want to systematically assess how many functional retinal channels are restored when we optogenetically target only RGCs types, and how this alteration to retinal processing changes the nature of responses in visual cortex. Methods: We used rd1 and Gnat1/2 blind mice. To perform optogenetics we intravitreally inject AAVs to express MW-opsin specifically in RGCs. To test visual function, we screen mice with a light-room/dark-room test. We also record RGCs activity on a 256-channel MEA array and present them different visual stimuli. To examine light responses in visual cortex, we perform in vivo 2-photon calcium imaging from V1. Results: We expressed MW-opsin to the retina of blind animals and found that their vision is restored when assessed with our light-room/dark-room screen. We obtained restored light responses from optogenetically-treated retinae and found these responses are all ON-type, with reduced diversity in orientation/direction selectivity and spatial/temporal frequency tuning. We have preliminary data from visual cortex, revealing single-cell light responses, with maintained presence of features such as orientation and direction selectivity. These results provide insights into the extent to which visual cortex can compensate for non-normal restored retinal responses.
The association between risk of obstructive sleep apnea and functional outcomes in individuals with Parkinson's disease within the Canadian Longitudinal Study on Aging.
2Translational Research in Respiratory Diseases, Research Institute of the McGill University Health Centre, Montreal, Quebec, Canada
3Department of Medicine and Department of Epidemiology, Biostatistics & Occupational Health, McGill University Health Centre, Montreal, Quebec, Canada
4Montreal Neurological Hospital, McGill University Centre, Montreal, Quebec, Canada
5Research Center, CIUSSS Nord-de-l'Ile-de-Montreal, Montreal, Quebec, Canada
6Department of Neurology and Neurosurgery, McGill University, Montreal General Hospital, Montreal, Quebec, Canada
7Respiratory Division and Sleep Laboratory, McGill University Health Centre, Montreal, Quebec, Canada
Corresponding Author: Teresa Gomes, email: teresa.gomes@mail.mcgill.ca
Abstract
Background: Obstructive sleep apnea (OSA) is frequently unnoticed when Parkinson’s disease (PD) motor symptoms are present and may result in functional limitations. This study aimed to assess the correlation between high risk for OSA and functional outcomes in PD individuals from a population cohort. Methods: PD individuals were identified in Canadian Longitudinal Study of Aging (CLSA) comprehensive cohort at baseline or at 3-year follow-up using a validated algorithm. High risk of OSA was determined at the corresponding timepoint using the STOP questionnaire (Loud Snoring, Tiredness/Sleepiness, Observed apneas, and high blood Pressure¬) >2. Functional measures included: Timed Up & Go, Standing Balance, Chair Rise, Four-Meter Walk test (gait) and Falls. Linear regression was performed to assess relationships between STOP scores and functional outcome, adjusted for potential confounders. Results: The findings showed that high STOP scores correlated with longer time to accomplish the Timed Up & Go and Four-Meter Walk tests, in unadjusted analyses and adjusting for age, gender, BMI, income, and education. Additionally, adjusted analyses revealed that in females only, higher STOP scores were linked to longer time to complete the Timed Up & Go and Four-Meter Walk test. Conclusion: High risk of OSA was associated with decreased mobility and physical function among PD participants in the CLSA population cohort. Results in stratified analyses were significant in female only. OSA may impact physical health differently depending on participants’ sex, or the STOP definition of OSA may identify different disease severity or characteristics between females and. males.
The Brain After COVID-19: Uncovering Evidence of Altered Brain Functions Through Electroencephalography
2Centre for Interdisciplinary Research in Rehabilitation of Greater Montreal, Montreal (CRIR), Canada.
3Department of Anesthesiology and Pain Medicine, Medicine, Université de Montréal.
4Centre for Advanced Research in Sleep Medicine, Hôpital du Sacré-Coeur de Montréal, CIUSSS du Nord-de-l'Île-de-Montréal, Montreal, Canada.
5School of Physical and Occupational Therapy, McGill University, Montreal, Canada.
6Centre for Outcomes Research and Evaluation, Research Institute of the McGill University Health Centre.
7Jewish Rehabilitation Hospital, CISSS-Laval, Laval, Canada.
Corresponding Author: Monserrat Casado Sánchez, Marie-Hélène Boudrias, email: monserrat.casado@mail.mcgill.ca, mh.boudrias@mcgill.ca
Abstract
A COVID-19 infection, though primarily a respiratory illness, can affect over 10 physiological systems, with over 200 reported symptoms. Around 30% of individuals infected with COVID-19 experience symptoms beyond 12 weeks post-infection, known as Post COVID-19 condition (PC19). The 3 main symptoms lingering 6 months post-infection include fatigue, brain fog, and post-exertional malaise (PEM). The brain is among the organs affected by COVID-19, where alteration in the blood-brain barrier, grey matter loss and decreased excitability have been reported. Thus, alterations in brain structure and functioning may underlie the residual and persisting PC19 symptomatology. This study utilizes electroencephalography (EEG) to explore the changes in brain oscillatory and functional properties during rest and hand movements between 20 PC19 individuals and 20 age-matched healthy participants. The aim is to identify the neural correlates associated with the 3 main lingering symptoms of PC19. Our main hypotheses are that compared to healthy individuals, PC19 individuals will report higher levels of fatigue, brain fog, and PEM, and these symptoms will be associated with abnormal brain patterns such as changes in syn/de-synchronization in the beta band during movement. Preliminary EEG results from 30 participants reveal decreased post-movement synchronization in the beta band. Behavioural data from clinical tests reflecting fatigue, brain fog, and PEM symptoms show significantly higher fatigue levels and cognitive complaints in the PC19 group. This study represents a critical step for identifying biomarkers of PC19, potentially aiding in developing precise diagnostic methods and targeted treatments to improve the debilitating symptoms experienced by this population.
The control costs of human brain dynamics
2Department of Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA, USA;
3Department of Neuroradiology, Klinikum rechts der Isar, TUM School of Medicine and Health, Technical University of Munich, Munich, Germany;
4Department of Neuroradiology, Uniklinikum Erlangen, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany;
5Research Group in Medical Imaging, SURA Ayudas Diagnósticas, Medellín, Colombia
Corresponding Author: Eric Gabriel Ceballos, email: eric.ceballosdominguez@mail.mcgill.ca
Abstract
The human brain is a complex system with high metabolic demands and extensive connectivity that requires control to balance energy consumption and functional efficiency over time. How this control is manifested on a whole-brain scale is largely unexplored, particularly what the associated costs are. Using network control theory, here we introduce a novel concept, time-averaged control energy (TCE), to quantify the cost of controlling human brain dynamics at rest, as measured from functional and diffusion MRI. Importantly, TCE spatially correlates with oxygen metabolism measures from positron emission tomography, providing insight into the bioenergetic footing of resting state control. Examining the temporal dimension of control costs, we find that brain state transitions along a hierarchical axis from sensory to association areas are more efficient in terms of control costs and more frequent within hierarchical groups than between. This inverse correlation between temporal control costs and state visits suggests a mechanism for maintaining functional diversity while minimizing energy expenditure. By unpacking the temporal dimension of control costs, we contribute to the neuroscientific understanding of how the brain governs its functionality while managing energy expenses.
The Efficacy of Intermittent Theta Burst Stimulation (iTBS) in Neural Facilitation and Motor Adaptation Learning
2Child Study Center, Yale School of Medicine, New Haven, Connecticut, USA
Corresponding Author: John Nicia, Ostry David, email: nicia.john@mail.mcgill.ca, david.ostry@mcgill.ca, david.ostry@yale.edu
Abstract
Intermittent theta burst stimulation (iTBS), a non-invasive brain stimulation (NIBS) technique based on the principles of theta-gamma coupling in the hippocampus, was developed to focally facilitate learning, memory, and cortical excitability in diverse brain regions. In work to date, the application of one block of iTBS produces highly inconsistent effects on behaviour and cortical activity. Thus, we explored the effects of two blocks of iTBS on motor learning and evoked potentials with aims of improving the efficacy of the technique. Two studies were conducted. In the first, participants received either one (iTBSx1) or two blocks (iTBSx2) of iTBS to either the primary motor cortex (M1), primary somatosensory cortex (S1), or a control area before training in an upper-limb force-field adaptation task, and measures of task consolidation and retention were obtained 24 hours after initial training. While iTBSx1 to S1 resulted in significantly impaired day one learning, no significant differences in learning or retention were observed following iTBSx2. In study two, changes in motor (MEP) and somatosensory evoked potentials (SEPs) were assessed post iTBSx2. In the MEP experiment, participants received brain stimulation to either M1 or S1, and no differences in MEP amplitudes were observed between conditions. In an analysis of SEPs obtained with EEG, however, iTBSx2 had a significant inhibitory effect on cortical activity. Overall, these findings demonstrate that iTBSx2 can increase interindividual response variability and does not improve the efficacy of the theta-burst protocol, suggesting a need for a more nuanced understanding of iTBS mechanisms.
The Impact of Screening for Anxious and Depressive Symptoms on the Outcome of Patients with a Mild Traumatic Brain Injury
2Department of Neurology and Neurosurgery, McGill University, Montreal, QC, Canada;
3Traumatic Brain Injury Program, McGill University Health Centre, Montreal, QC, Canada
Corresponding Author: Maude Laguë-Beauvais, Judith Marcoux, email: maude.lague-beauvais@muhc.mcgill.ca, judith.marcoux@mcgill.ca
Abstract
Approximately 27-69 million individuals worldwide sustain a mild Traumatic Brain Injury (mTBI) each year, making it an important public health concern. Many victims experience post-injury anxiety and/or depression, which are associated with more post-concussive symptoms and worse functional outcomes. We sought to determine a systematic process to document the presence of symptoms of anxiety and depression in mTBI patients to prevent negative impacts on their recovery. We thus administered the Generalized Anxiety Disorder-7 (GAD-7) and Center for Epidemiologic Studies Depression Scale Revised (CESDR-10) questionnaires, no more than three months after injury, to screen for these symptoms. A retrospective chart review was performed for 328 patients from the Montreal General Hospital mTBI Clinic who either received these questionnaires (N=143, Mage=40.36, SDage=15.557, Nfemale=90, Nmale=53) or did not (N=185, Mage=41.17, SDage=16.449, Nfemale=114, Nmale=71). We documented the total number of referrals to various interventions that each patient received during their clinic visits, quantified patient outcomes using the Glasgow Outcome Scale-Extended (GOS-E) and compared these variables between groups using ANOVA. Patients who received the questionnaires (M=1.34, SD=0.978) were referred to significantly more interventions than those who did not (M=0.90, SD=0.876, p<0.001) and the rate of referral positively correlated with GAD-7 and CESDR-10 scores. These patients also have significantly higher GOS-E scores (M=7.55, SD=0.636) than those who did not receive the questionnaires (M=7.15, SD=0.595, p<0.001), indicating better functional outcomes. Therefore, screening for symptoms of anxiety and depression post mTBI helps clinicians refer patients to the appropriate resources, which facilitates recovery and improves patient outcomes.
The N400 effect in school-aged children with and without a reading disability: evidence for a relationship with Set-for-Variability (SfV) and reading ability.
2Faculty of Education, York University, Toronto, ON, Canada
Corresponding Author: Badriah Basma , email: badriah.basma@mail.mcgill.ca
Abstract
Background. The N400 ERP demonstrated to be an index of lexical-semantic processing, can be used to assess the neural underpinnings of reading disability (RD). We aimed to assess (i) whether a N400 differs between children with and without RD, and (ii) if N400 potentials are correlated with Set-for-Variability (SfV), a measure of children’s ability to generate alternative pronunciation when they read unknown words, and other reading abilities. Methods. Twenty neurotypical (NT) children and 31 9-year-old children with RD completed a sentence-judgement task (congruent and incongruent conditions) while N400 was measured; SfV and a battery of standardized literacy tests were also completed. Results. A relatively larger N400 negative peak amplitude was found for incongruent sentences for the TD group. SfV was negatively correlated with the N400 latency in the RD group only, while word reading measures were negatively associated with onset latency in both RD and TD groups. Conclusion. Children with RD demonstrated aberrant N400 profiles alongside delays in SfV, with the N400 profiles associated with SfV ability. Results support the N400 as a physiological index word reading abilities in young learners with RD. Action. Results will serve as the basis for reading intervention studies using electrophysiological (N400) and behavioural (SfV) outcome measures.
Training at-home on a dichoptic reading application to improve vision in adults with amblyopia
2Department of Ophthalmology and Visual Sciences, McGill University, Montreal, QC, Canada
3Research Institute of the McGill University Health Centre, Montreal, QC, Canada
Corresponding Author: Nicole Dranitsaris, email: nicole.dranitsaris@mail.mcgill.ca
Abstract
Current research on amblyopia treatment methods has shifted focus from typical patching, which is only applicable in childhood, to using dichoptic tasks. This study leveraged a new approach implicating a daily task, reading, to improve amblyopic vision. Here, we assessed if at-home training on a dichoptic e-book reading application (DEBRA) can be an alternative treatment for binocular vision in amblyopia. DEBRA was uploaded onto tablets displaying E-books with words in red/green/black presentation. Anaglyph red/green glasses allowed different text to be shown to each eye simultaneously, forcing the individual to combine the input from both eyes. Adult amblyopic participants brought the technology home and trained for one hour/day for two weeks. Before and after the training participants were given a full visual assessment, then reading speed and eye movements patterns while reading were recorded. After training they also completed a visual comfort questionnaire. Results demonstrated significantly improved monocular visual acuity, in both amblyopic and fellow eyes. Amblyopic participants also presented reading gaze patterns closer to those of controls following the training. Based on visual comfort questionnaire responses, majority of participants did not experience visual discomfort while completing the dichoptic training. This study demonstrated that daily training on a dichoptic reading application at-home for 14 hours improves amblyopic visual function. More data will clarify if eye movement patterns and other altered ophthalmic factors in amblyopia can be treated by completing the training. Future steps are aimed at collecting more data from amblyopes and ameliorating the user-friendliness of the application.
Using Transcranial Opto-stimulation to Investigate the Link Between Early-Life Seizures and Social Behaviour
Corresponding Author: Haleigh Bach, email: haleigh.bach@mail.mcgill.ca
Abstract
Autism and childhood epilepsy are co-morbid neurodevelopmental disorders with a complex and unexplained etiology. Early-life seizures (ELS) are associated with epilepsy development and lead to sociability deficits, a symptom of ASD, in rodent models. However, rodent studies use inconsistent ELS models and lack investigation into the ELS-induced circuitry alterations leading to sociability deficits. We are developing a non-invasive ELS model that harnesses the sensitivity of the ChRmine opsin for unparalleled spatiotemporal precision of ELS induction. We have optimized a transverse injection procedure on P0-P1 mice to express ChRmine in the cortex and hippocampus. Then, we piloted transcranial laser stimulation to activate ChRmine during the maturation of the cortico-striatal pathway important for social behaviour in mice. Injections of pentylenetetrazol (PTZ), were conducted to serve as a positive control, and cFos staining will be used to compare cortical activation between our optogenetic model and PTZ. Alterations in social behaviour will be assessed, and compared, across the lifespan on P30 and P60 in both PTZ-injected and laser-stimulated mice. We will be presenting behavioural data of two cohorts of mice injected with PTZ at different windows of cortico-striatal development, with the goal of zooming in on a putative period critical for the development of mouse social behaviour. The results of this project will be useful in creating a non-invasive and spatiotemporally precise ELS model to uncover the neural circuitry underlying ELS-induced sociability deficits.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.