Publications

Hope Center Member Publications: October 22, 2023

Deep representation learning identifies associations between physical activity and sleep patterns during pregnancy and prematurity” (2023) npj Digital Medicine

Deep representation learning identifies associations between physical activity and sleep patterns during pregnancy and prematurity
(2023) npj Digital Medicine, 6 (1), art. no. 171, . 

Ravindra, N.G.a b c , Espinosa, C.a b c , Berson, E.a c d , Phongpreecha, T.a c d , Zhao, P.e f , Becker, M.a b c , Chang, A.L.a b c , Shome, S.a b c , Marić, I.a b c , De Francesco, D.a b c , Mataraso, S.a b c , Saarunya, G.a b c , Thuraiappah, M.a b c , Xue, L.a b c , Gaudillière, B.a , Angst, M.S.a , Shaw, G.M.b , Herzog, E.D.e , Stevenson, D.K.b , England, S.K.f , Aghaeepour, N.a b c

a Department of Anesthesiology, Perioperative and Pain Medicine, Stanford School of Medicine, Stanford, CA, United States
b Department of Pediatrics, Stanford School of Medicine, Stanford, CA, United States
c Department of Biomedical Data Science, Stanford University, Stanford, CA, United States
d Department of Pathology, Stanford School of Medicine, Stanford, CA, United States
e Department of Biology, Washington University in St. Louis, St. Louis, MO, United States
f Department of Obstetrics and Gynecology, Washington University in St. Louis, St. Louis, MO, United States

Abstract
Preterm birth (PTB) is the leading cause of infant mortality globally. Research has focused on developing predictive models for PTB without prioritizing cost-effective interventions. Physical activity and sleep present unique opportunities for interventions in low- and middle-income populations (LMICs). However, objective measurement of physical activity and sleep remains challenging and self-reported metrics suffer from low-resolution and accuracy. In this study, we use physical activity data collected using a wearable device comprising over 181,944 h of data across N = 1083 patients. Using a new state-of-the art deep learning time-series classification architecture, we develop a ‘clock’ of healthy dynamics during pregnancy by using gestational age (GA) as a surrogate for progression of pregnancy. We also develop novel interpretability algorithms that integrate unsupervised clustering, model error analysis, feature attribution, and automated actigraphy analysis, allowing for model interpretation with respect to sleep, activity, and clinical variables. Our model performs significantly better than 7 other machine learning and AI methods for modeling the progression of pregnancy. We found that deviations from a normal ‘clock’ of physical activity and sleep changes during pregnancy are strongly associated with pregnancy outcomes. When our model underestimates GA, there are 0.52 fewer preterm births than expected (P = 1.01e − 67, permutation test) and when our model overestimates GA, there are 1.44 times (P = 2.82e − 39, permutation test) more preterm births than expected. Model error is negatively correlated with interdaily stability (P = 0.043, Spearman’s), indicating that our model assigns a more advanced GA when an individual’s daily rhythms are less precise. Supporting this, our model attributes higher importance to sleep periods in predicting higher-than-actual GA, relative to lower-than-actual GA (P = 1.01e − 21, Mann-Whitney U). Combining prediction and interpretability allows us to signal when activity behaviors alter the likelihood of preterm birth and advocates for the development of clinical decision support through passive monitoring and exercise habit and sleep recommendations, which can be easily implemented in LMICs. © 2023, Springer Nature Limited.

Funding details
National Institutes of HealthNIH19PABHI34580007, 1R01HL139844, P01HD106414, P30AG066515, R01AG058417, R01HD105256, R35GM138353, R61NS114926
Burroughs Wellcome FundBWF1019816
Bill and Melinda Gates FoundationBMGFINV-001734, INV-003225, OPP1113682
March of Dimes FoundationMDF
American Heart AssociationAHA19PAB / HI34580007
Robertson Foundation

Document Type: Article
Publication Stage: Final
Source: Scopus

A novel immune modulator IM33 mediates a glia-gut-neuronal axis that controls lifespan” (2023) Neuron

A novel immune modulator IM33 mediates a glia-gut-neuronal axis that controls lifespan
(2023) Neuron, 111 (20), pp. 3244-3254.e8. 

Xu, W.a b , Rustenhoven, J.a b c d , Nelson, C.A.b , Dykstra, T.a b , Ferreiro, A.e f , Papadopoulos, Z.a b g , Burnham, C.-A.D.b , Dantas, G.b e f h , Fremont, D.H.b h i j , Kipnis, J.a b g

a Brain Immunology and Glia (BIG) Center, Washington University in St. Louis, School of Medicine, St. Louis, MO, United States
b Department of Pathology and Immunology, Washington University in St. Louis, School of Medicine, St. Louis, MO, United States
c Department of Pharmacology and Clinical Pharmacology, The University of Auckland, Auckland, New Zealand
d Centre for Brain Research, The University of Auckland, Auckland, New Zealand
e The Edison Family Center for Genome Sciences and Systems Biology, Washington University in St. Louis, School of Medicine, St. Louis, MO, United States
f Department of Biomedical Engineering, Washington University in St. Louis, St. Louis, MO, United States
g Neuroscience Graduate Program, School of Medicine, Washington University in St. Louis, School of Medicine, St. Louis, MO, United States
h Department of Molecular Microbiology, Washington University in St. Louis, School of Medicine, St. Louis, MO, United States
i The Andrew M. and Jane M. Bursky Center for Human Immunology and Immunotherapy Programs, Washington University School of Medicine, St. Louis, MO, United States
j Department of Biochemistry & Molecular Biophysics, Washington University School of Medicine, St. Louis, MO, United States

Abstract
Aging is a complex process involving various systems and behavioral changes. Altered immune regulation, dysbiosis, oxidative stress, and sleep decline are common features of aging, but their interconnection is poorly understood. Using Drosophila, we discover that IM33, a novel immune modulator, and its mammalian homolog, secretory leukocyte protease inhibitor (SLPI), are upregulated in old flies and old mice, respectively. Knockdown of IM33 in glia elevates the gut reactive oxygen species (ROS) level and alters gut microbiota composition, including increased Lactiplantibacillus plantarum abundance, leading to a shortened lifespan. Additionally, dysbiosis induces sleep fragmentation through the activation of insulin-producing cells in the brain, which is mediated by the binding of Lactiplantibacillus plantarum-produced DAP-type peptidoglycan to the peptidoglycan recognition protein LE (PGRP-LE) receptor. Therefore, IM33 plays a role in the glia-microbiota-neuronal axis, connecting neuroinflammation, dysbiosis, and sleep decline during aging. Identifying molecular mediators of these processes could lead to the development of innovative strategies for extending lifespan. © 2023 Elsevier Inc.

Author Keywords
brain-gut axis;  Drosophila;  glia-gut axis;  IM33;  longevity;  neuroimmunology

Funding details
R01-AT009741
National Institutes of HealthNIHP40OD018537
Washington University in St. LouisWUSTL
University of VirginiaUV
National Center for Complementary and Integrative HealthNCCIHDP1AT010416

Document Type: Article
Publication Stage: Final
Source: Scopus

Biallelic variants in COQ7 cause distal hereditary motor neuropathy with upper motor neuron signs” (2023) Brain: A Journal of Neurology

Biallelic variants in COQ7 cause distal hereditary motor neuropathy with upper motor neuron signs
(2023) Brain: A Journal of Neurology, 146 (10), pp. 4191-4199. 

Rebelo, A.P.a , Tomaselli, P.J.b , Medina, J.a , Wang, Y.c , Dohrn, M.F.a d , Nyvltova, E.e , Danzi, M.C.a , Garrett, M.f , Smith, S.E.f , Pestronk, A.f , Li, C.f , Ruiz, A.a , Jacobs, E.a , Feely, S.M.E.g , França, M.C.b , Gomes, M.V.b , Santos, D.F.h , Kumar, S.i , Lombard, D.B.i , Saporta, M.a , Hekimi, S.c , Barrientos, A.e , Weihl, C.f , Shy, M.E.g , Marques, W.b , Zuchner, S.a

a Dr. John T. Macdonald Foundation, Department of Human Genetics and John P. Hussman Institute for Human Genomics, University of Miami, Miller School of Medicine, Miami, FL 33136, United States
b Department of Neurology, University of São Paulo, Ribeirão Preto, 14048-900, Brazil
c Department of Biology, McGill University, Montreal, QC H3A 1A1, Canada
d Department of Neurology, Medical Faculty, RWTH Aachen University, Aachen, 52074, Germany
e Department of Neurology, University of Miami Miller School of Medicine, Miami, FL 33136, United States
f Department of Neurology, Washington University, St. Louis, United States
g Department of Neurology, Carver College of Medicine, University of Iowa, Iowa City, IA 52242, United States
h Department of Neurology, Federal University of Uberlândia, Uberlândia, MG 38405-320, Brazil
i Department of Pathology & Laboratory Medicine, Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, FL 33136, United States

Abstract
COQ7 encodes a hydroxylase responsible for the penultimate step of coenzyme Q10 (CoQ10) biosynthesis in mitochondria. CoQ10 is essential for multiple cellular functions, including mitochondrial oxidative phosphorylation, lipid metabolism, and reactive oxygen species homeostasis. Mutations in COQ7 have been previously associated with primary CoQ10 deficiency, a clinically heterogeneous multisystemic mitochondrial disorder. We identified COQ7 biallelic variants in nine families diagnosed with distal hereditary motor neuropathy with upper neuron involvement, expending the clinical phenotype associated with defects in this gene. A recurrent p.Met1? change was identified in five families from Brazil with evidence of a founder effect. Fibroblasts isolated from patients revealed a substantial depletion of COQ7 protein levels, indicating protein instability leading to loss of enzyme function. High-performance liquid chromatography assay showed that fibroblasts from patients had reduced levels of CoQ10, and abnormal accumulation of the biosynthetic precursor DMQ10. Accordingly, fibroblasts from patients displayed significantly decreased oxygen consumption rates in patients, suggesting mitochondrial respiration deficiency. Induced pluripotent stem cell-derived motor neurons from patient fibroblasts showed significantly increased levels of extracellular neurofilament light protein, indicating axonal degeneration. Our findings indicate a molecular pathway involving CoQ10 biosynthesis deficiency and mitochondrial dysfunction in patients with distal hereditary motor neuropathy. Further studies will be important to evaluate the potential benefits of CoQ10 supplementation in the clinical outcome of the disease. © The Author(s) 2023. Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Author Keywords
charcot-Marie-Tooth disease;  coQ10;  hereditary motor neuropathy;  mitochondria;  motor neuron

Document Type: Article
Publication Stage: Final
Source: Scopus

Peripheral CCL2-CCR2 signalling contributes to chronic headache-related sensitization” (2023) Brain: A Journal of Neurology

Peripheral CCL2-CCR2 signalling contributes to chronic headache-related sensitization
(2023) Brain: A Journal of Neurology, 146 (10), pp. 4274-4291. 

Ryu, S., Liu, X., Guo, T., Guo, Z., Zhang, J., Cao, Y.-Q.

Department of Anesthesiology and Washington University Pain Center, Washington University School of Medicine, Campus Box MSC 8054-86-05, St. Louis, MO 63110, United States

Abstract
Migraine, especially chronic migraine, is highly debilitating and still lacks effective treatment. The persistent headache arises from activation and sensitization of primary afferent neurons in the trigeminovascular pathway, but the underlying mechanisms remain incompletely understood. Animal studies indicate that signalling through chemokine C-C motif ligand 2 (CCL2) and C-C motif chemokine receptor 2 (CCR2) mediates the development of chronic pain after tissue or nerve injury. Some migraine patients had elevated CCL2 levels in CSF or cranial periosteum. However, whether the CCL2-CCR2 signalling pathway contributes to chronic migraine is not clear. Here, we modelled chronic headache with repeated administration of nitroglycerin (NTG, a reliable migraine trigger in migraineurs) and found that both Ccl2 and Ccr2 mRNA were upregulated in dura and trigeminal ganglion (TG) tissues that are implicated in migraine pathophysiology. In Ccl2 and Ccr2 global knockout mice, repeated NTG administration did not evoke acute or persistent facial skin hypersensitivity as in wild-type mice. Intraperitoneal injection of CCL2 neutralizing antibodies inhibited chronic headache-related behaviours induced by repeated NTG administration and repetitive restraint stress, suggesting that the peripheral CCL2-CCR2 signalling mediates headache chronification. We found that CCL2 was mainly expressed in TG neurons and cells associated with dura blood vessels, whereas CCR2 was expressed in subsets of macrophages and T cells in TG and dura but not in TG neurons under both control and disease states. Deletion of Ccr2 gene in primary afferent neurons did not alter NTG-induced sensitization, but eliminating CCR2 expression in either T cells or myeloid cells abolished NTG-induced behaviours, indicating that both CCL2-CCR2 signalling in T cells and macrophages are required to establish chronic headache-related sensitization. At cellular level, repeated NTG administration increased the number of TG neurons that responded to calcitonin-gene-related peptide (CGRP) and pituitary adenylate cyclase activating polypeptide (PACAP) as well as the production of CGRP in wild-type but not Ccr2 global knockout mice. Lastly, co-administration of CCL2 and CGRP neutralizing antibodies was more effective in reversing NTG-induced behaviours than individual antibodies. Taken together, these results suggest that migraine triggers activate CCL2-CCR2 signalling in macrophages and T cells. This consequently enhances both CGRP and PACAP signalling in TG neurons, ultimately leading to persistent neuronal sensitization underlying chronic headache. Our work not only identifies the peripheral CCL2 and CCR2 as potential targets for chronic migraine therapy, but also provides proof-of-concept that inhibition of both peripheral CGRP and CCL2-CCR2 signalling is more effective than targeting either pathway alone. © The Author(s) 2023. Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Author Keywords
CCL2;  CCR2;  headache;  migraine;  neuroimmune interaction;  peripheral sensitization

Document Type: Article
Publication Stage: Final
Source: Scopus

Incorporating physics to overcome data scarcity in predictive modeling of protein function: A case study of BK channels” (2023) PLoS Computational Biology

Incorporating physics to overcome data scarcity in predictive modeling of protein function: A case study of BK channels
(2023) PLoS Computational Biology, 19 (9), art. no. e1011460, . 

Nordquist, E.a , Zhang, G.b , Barethiya, S.a , Ji, N.c , White, K.M.b , Han, L.b , Jia, Z.a , Shi, J.b , Cui, J.b , Chen, J.a

a Department of Chemistry, University of Massachusetts Amherst, Amherst, MA, United States
b Department of Biomedical Engineering, Center for the Investigation of Membrane Excitability Disorders, Cardiac Bioelectricity and Arrhythmia Center, Washington University in St. Louis, St. Louis, MO, United States
c Department of Biology, Boston College, Chestnut Hill, MA, United States

Abstract
Machine learning has played transformative roles in numerous chemical and biophysical problems such as protein folding where large amount of data exists. Nonetheless, many important problems remain challenging for data-driven machine learning approaches due to the limitation of data scarcity. One approach to overcome data scarcity is to incorporate physical principles such as through molecular modeling and simulation. Here, we focus on the big potassium (BK) channels that play important roles in cardiovascular and neural systems. Many mutants of BK channel are associated with various neurological and cardiovascular diseases, but the molecular effects are unknown. The voltage gating properties of BK channels have been characterized for 473 site-specific mutations experimentally over the last three decades; yet, these functional data by themselves remain far too sparse to derive a predictive model of BK channel voltage gating. Using physics-based modeling, we quantify the energetic effects of all single mutations on both open and closed states of the channel. Together with dynamic properties derived from atomistic simulations, these physical descriptors allow the training of random forest models that could reproduce unseen experimentally measured shifts in gating voltage, ΔV1/2, with a RMSE ~ 32 mV and correlation coefficient of R ~ 0.7. Importantly, the model appears capable of uncovering nontrivial physical principles underlying the gating of the channel, including a central role of hydrophobic gating. The model was further evaluated using four novel mutations of L235 and V236 on the S5 helix, mutations of which are predicted to have opposing effects on V1/2 and suggest a key role of S5 in mediating voltage sensor-pore coupling. The measured ΔV1/2 agree quantitatively with prediction for all four mutations, with a high correlation of R = 0.92 and RMSE = 18 mV. Therefore, the model can capture nontrivial voltage gating properties in regions where few mutations are known. The success of predictive modeling of BK voltage gating demonstrates the potential of combining physics and statistical learning for overcoming data scarcity in nontrivial protein function prediction. Copyright: © 2023 Nordquist et al.

Funding details
T32 GM008515, T32 GM139789
National Institutes of HealthNIHR01 HL070393, R01 HL142301, R35 GM144045, RO1 NS060706

Document Type: Article
Publication Stage: Final
Source: Scopus

Toward peripheral nerve mechanical characterization using Brillouin imaging spectroscopy” (2023) Neurophotonics

Toward peripheral nerve mechanical characterization using Brillouin imaging spectroscopy
(2023) Neurophotonics, 10 (3), art. no. 035007, . Cited 1 time.

Cheburkanov, V.a , Du, J.b c , Brogan, D.M.d , Berezin, M.Y.b c , Yakovlev, V.V.a

a Texas A&M University, Department of Biomedical Engineering, College Station, TX, United States
b Washington University School of Medicine, Department of Radiology, St. Louis, MO, United States
c Washington University, Institute of Materials Science and Engineering, St. Louis, MO, United States
d Washington University School of Medicine, Department of Orthopedic Surgery, St. Louis, MO, United States

Abstract
Significance: Peripheral nerves are viscoelastic tissues with unique elastic characteristics. Imaging of peripheral nerve elasticity is important in medicine, particularly in the context of nerve injury and repair. Elasticity imaging techniques provide information about the mechanical properties of peripheral nerves, which can be useful in identifying areas of nerve damage or compression, as well as assessing the success of nerve repair procedures. Aim: We aim to assess the feasibility of Brillouin microspectroscopy for peripheral nerve imaging of elasticity, with the ultimate goal of developing a new diagnostic tool for peripheral nerve injury in vivo. Approach: Viscoelastic properties of the peripheral nerve were evaluated with Brillouin imaging spectroscopy. Results: An external stress exerted on the fixed nerve resulted in a Brillouin shift. Quantification of the shift enabled correlation of the Brillouin parameters with nerve elastic properties. Conclusions: Brillouin microscopy provides sufficient sensitivity to assess viscoelastic properties of peripheral nerves. © The Authors.

Author Keywords
brillouin;  confocal imaging;  ex vivo;  nerve

Funding details
National Science FoundationNSFCMMI-1826078
National Institutes of HealthNIHR01GM127696, R21CA269099, R21GM142107
U.S. Food and Drug AdministrationFDA80ARC023CA002, R01 CA208623
National Aeronautics and Space AdministrationNASA
Air Force Office of Scientific ResearchAFOSRFA9550-20-1-0366, FA9550-20-1-0367
Army Research LaboratoryARLW911NF-17-2-0144
Biomedical Advanced Research and Development AuthorityBARDA

Document Type: Article
Publication Stage: Final
Source: Scopus

Longitudinal predictors of health-related quality of life in isolated dystonia” (2023) Journal of Neurology

Longitudinal predictors of health-related quality of life in isolated dystonia
(2023) Journal of Neurology, . 

Junker, J.a b , Hall, J.c , Berman, B.D.d , Vidailhet, M.e f , Roze, E.e , Bäumer, T.g , Malaty, I.A.h , Shukla, A.W.h , Jankovic, J.i , Reich, S.G.j , Espay, A.J.k , Duque, K.R.k , Patel, N.l , Perlmutter, J.S.m , Jinnah, H.A.n , Brandt, V.o , Brüggemann, N.a b

a Department of Neurology, University of Luebeck, Ratzeburger Allee 160, SH, Lübeck, 23538, Germany
b Institute of Neurogenetics, University of Luebeck, Luebeck, Germany
c Southampton Education School, University of Southampton, Southampton, United Kingdom
d Department of Neurology, Virginia Commonwealth University, Richmond, VA, United States
e Departement de Neurologie, AP-HP, Hopital de La Pitie-Salpetriere, Paris, France
f Institut du Cerveau_ Paris Brain Institute-ICM, INSERM 1127, CNRS 7225, Sorbonne Université, Paris, France
g Institute of Systems Motor Science, University of Luebeck, Luebeck, Germany
h Department of Neurology, Fixel Institute for Neurologic Disorders, University of Florida, Gainesville, FL, United States
i Parkinson’s Disease Center and Movement Disorders Clinic, Department of Neurology, Baylor College of Medicine, Houston, TX, United States
j Department of Neurology, School of Medicine, University of Maryland, Baltimore, MD, United States
k Department of Neurology, University of Cincinnati, Cincinnati, OH, United States
l RUSH Parkinson’s Disease and Movement Disorders Center, Department of Neurological Science, RUSH University Medical Center Chicago, Chicago, IL, United States
m Departments of Neurology, Radiology and Neuroscience, Washington University in St. Louis, St. Louis, MO, United States
n Department of Neurology and Human Genetics, Emory University, Atlanta, GA, United States
o School of Psychology, Centre for Innovation in Mental Health, University of Southampton, Southampton, United Kingdom

Abstract
Objective: To determine longitudinal predictors of health-related quality of life (HR-QoL) in an international multicenter cohort of patients with isolated dystonia. Methods: Out of 603 dystonia patients prospectively enrolled in the Natural History Dystonia Coalition study, 155 were assessed three times within 2 years for HR-QoL, symptoms of depression, generalized anxiety disorder (GAD), and social anxiety disorder (SAD), as well as dystonia severity and dystonic tremor. In addition, the impact of botulinum neurotoxin (BoNT) injections on HR-QoL was evaluated after 1 year. Results: Depressive symptoms at baseline predicted lower HR-QoL on all subscales after 2 years (all p ≤ 0.001). Higher GAD scores at baseline predicted lower HR-QoL related to general health, pain and emotional well-being, whereas higher SAD scores predicted higher pain-related QoL after 2 years (all p ≤ 0.006). Dystonia severity at baseline predicted social functioning (p = 0.002). Neither dystonic tremor, age, or sex predicted HR-QoL at 2 years. Two latent categories were revealed across the three-time points: Category 1 with higher total HR-QoL scores (mean HR-QoL = 74.4% ± 16.1), susceptible to symptoms of depression and SAD, and Category 2 with lower total HR-QoL scores (mean HR-QoL = 45.5% ± 17.6), susceptible to symptoms of GAD. HR-QoL improved over the course of 1 year irrespective of the use of BoNT. Conclusion: The longitudinal impact of psychiatric symptoms on HR-QoL emphasizes the importance of incorporating mental health treatment, in particular also the therapy of anxiety disorders, into treatment regimens for dystonia. © 2023, The Author(s).

Author Keywords
Anxiety;  BoNT;  Depression;  Dystonia;  Quality of life

Funding details
U54 TR001456
National Institute of Neurological Disorders and StrokeNINDSU54 NS065701, U54 NS116025
Dystonia CoalitionDC
Rare Diseases Clinical Research NetworkRDCRN

Document Type: Article
Publication Stage: Article in Press
Source: Scopus

Geroscience-Centric Perspective for Geriatric Psychiatry: Integrating Aging Biology With Geriatric Mental Health Research” (2023) American Journal of Geriatric Psychiatry

Geroscience-Centric Perspective for Geriatric Psychiatry: Integrating Aging Biology With Geriatric Mental Health Research
(2023) American Journal of Geriatric Psychiatry, . 

Diniz, B.S.a , Seitz-Holland, J.b c , Sehgal, R.d , Kasamoto, J.d , Higgins-Chen, A.T.e f , Lenze, E.g

a UConn Center on Aging & Department of Psychiatry (BSD), School of Medicine, University of Connecticut Health Center, Farmington, CT, United States
b Department of Psychiatry (JSH), Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, United States
c Department of Psychiatry (JSH), Massachusetts General Hospital, Harvard Medical School, Boston, MA, United States
d Program in Computational Biology and Bioinformatics (RS, JK), Yale University, New Haven, CT, United States
e Department of Psychiatry (ATHC), Yale University School of Medicine, New Haven, CT, United States
f Department of Pathology (ATHC), Yale University School of Medicine, New Haven, CT, United States
g Department of Psychiatry (EL), School of Medicine, Washington University at St. Louis, St. Louis, MO, United States

Abstract
The geroscience hypothesis asserts that physiological aging is caused by a small number of biological pathways. Despite the explosion of geroscience research over the past couple of decades, the research on how serious mental illnesses (SMI) affects the biological aging processes is still in its infancy. In this review, we aim to provide a critical appraisal of the emerging literature focusing on how we measure biological aging systematically, and in the brain and how SMIs affect biological aging measures in older adults. We will also review recent developments in the field of cellular senescence and potential targets for interventions for SMIs in older adults, based on the geroscience hypothesis. © 2023 American Association for Geriatric Psychiatry

Author Keywords
Aging;  geroscience;  late-life depression;  serious mental illnesses

Funding details
National Institutes of HealthNIHR01MH115953
Brain and Behavior Research FoundationBBRF
Merck
Patient-Centered Outcomes Research InstitutePCORI
University of WashingtonUWR01AG057912, R01AG068937, UL1TR002345
Janssen Pharmaceuticals

Document Type: Article
Publication Stage: Article in Press
Source: Scopus

Towards Outcome-Driven Patient Subgroups: A Machine Learning Analysis Across Six Depression Treatment Studies” (2023) American Journal of Geriatric Psychiatry

Towards Outcome-Driven Patient Subgroups: A Machine Learning Analysis Across Six Depression Treatment Studies
(2023) American Journal of Geriatric Psychiatry, . 

Benrimoh, D.a b c , Kleinerman, A.d , Furukawa, T.A.e , III, C.F.R.f g , Lenze, E.J.h , Karp, J.i , Mulsant, B.j , Armstrong, C.c , Mehltretter, J.c , Fratila, R.c , Perlman, K.a c , Israel, S.c , Popescu, C.c , Golden, G.c , Qassim, S.c , Anacleto, A.c , Tanguay-Sela, M.c , Kapelner, A.k , Rosenfeld, A.d , Turecki, G.a

a Department of Psychiatry (DB, KP, GT), McGill University, Montreal, Canada
b Department of Psychiatry (DB), Stanford University, Stanford, CA, United States
c Aifred Health (DB, CA, JM, RF, KP, SI, CP, GG, SQ, AA, MTS), Montreal, Canada
d Bar-Ilan University (AK, AR), Ramat Gan, Israel
e Department of Health Promotion and Human Behavior (TAF), Kyoto University Graduate School of Medicine/School of Public Health, Japan
f Department of Psychiatry (CFR), University of Pittsburgh School of Medicine, United States
g Department of Psychiatry (CFR), Tufts University School of Medicine, United States
h Department of Psychiatry (EJL), Washington University School of Medicine, St. Louis, MS, United States
i Department of Psychiatry (JK), University of Arizona, United States
j Department of Psychiatry (BM), University of Toronto, Canada
k Department of Mathematics (AK), Queens College, CUNY, United States

Abstract
Background: Major depressive disorder (MDD) is a heterogeneous condition; multiple underlying neurobiological and behavioral substrates are associated with treatment response variability. Understanding the sources of this variability and predicting outcomes has been elusive. Machine learning (ML) shows promise in predicting treatment response in MDD, but its application is limited by challenges to the clinical interpretability of ML models, and clinicians often lack confidence in model results. In order to improve the interpretability of ML models in clinical practice, our goal was to demonstrate the derivation of treatment-relevant patient profiles comprised of clinical and demographic information using a novel ML approach. Methods: We analyzed data from six clinical trials of pharmacological treatment for depression (total n = 5438) using the Differential Prototypes Neural Network (DPNN), a ML model that derives patient prototypes which can be used to derive treatment-relevant patient clusters while learning to generate probabilities for differential treatment response. A model classifying remission and outputting individual remission probabilities for five first-line monotherapies and three combination treatments was trained using clinical and demographic data. Prototypes were evaluated for interpretability by assessing differences in feature distributions (e.g. age, sex, symptom severity) and treatment-specific outcomes. Results: A 3-prototype model achieved an area under the receiver operating curve of 0.66 and an expected absolute improvement in remission rate for those receiving the best predicted treatment of 6.5% (relative improvement of 15.6%) compared to the population remission rate. We identified three treatment-relevant patient clusters. Cluster A patients tended to be younger, to have increased levels of fatigue, and more severe symptoms. Cluster B patients tended to be older, female, have less severe symptoms, and the highest remission rates. Cluster C patients had more severe symptoms, lower remission rates, more psychomotor agitation, more intense suicidal ideation, and more somatic genital symptoms. Conclusion: It is possible to produce novel treatment-relevant patient profiles using ML models; doing so may improve interpretability of ML models and the quality of precision medicine treatments for MDD. © 2023

Author Keywords
artificial intelligence;  machine learning;  major depression;  subgroups

Funding details
2153, 2158
2199
2148
National Institutes of HealthNIH
National Institute of Mental HealthNIMH
American Association for Geriatric PsychiatryAAGP
University of Pittsburgh
University of MarylandUMD
University of South FloridaUSF
National Research Council CanadaNRC
Shionogi2020-548587, 2022-082495
Kyoto University

Document Type: Article
Publication Stage: Article in Press
Source: Scopus