39339 research outputs found
Sort by
Evaluating vitamin C-related gene-environment and metabolite-environment interaction effects on intraocular pressure in the Canadian Longitudinal Study on Aging
No full textAbstract High intraocular pressure (IOP) is an important risk factor for glaucoma, which is influenced by genetic and environmental factors. However, the etiology of high IOP remains uncertain. Metabolites are compounds involved in metabolism which provide a link between the internal (genetic) and external environments. O-methylascorbate has been reported to be associated with IOP. In addition, researchers have identified several genetic variants which are associated with metabolite concentrations, including O-methylascorbate and another vitamin C related metabolite, ascorbic acid 2-sulfate. We aimed to understand how O-methylascorbate and ascorbic acid 2-sulfate, or genetic variants associated with these metabolites, modify the associations between dietary environmental variables and IOP. We used data from 8060 participants of the Canadian Longitudinal Study on Aging. Using linear models adjusted for relevant covariates, we tested for interactions between six genetic variants previously found to be associated with O-methylascorbate and ascorbic acid 2-sulfate and four environmental variables related to diet (alcohol consumption frequency, smoking status, fruit consumption, and vegetable consumption). We also tested for interactions between serum concentrations of O-methylascorbate and ascorbic acid 2-sulfate and these environmental factors. We used a False Discovery Rate approach to correct for the 32 interaction tests performed. One interaction was suggestively significant after multiple testing correction (adjusted P-value < 0.1): rs8050812 and alcohol consumption frequency. Understanding how genetic variants and metabolites interact with the environment could shed light on biological pathways controlling IOP and lead to improved prevention and treatment of glaucoma
Crosstalk Between Mesenchymal Stem/Stromal Cells and Host Immune Cells as a Critical Mechanism of Immunomodulation in Severe Viral Lung Infection and Sepsis
No full textSepsis is characterized by a systemic dysregulated immune response to severe infection. Research has shown that mesenchymal stromal/stem cells (MSCs) exert immunemodulatory functions in vitro and therapeutic benefits in various animal models of severe immune diseases. These promising results have driven the translation of MSCs as cellular therapy into clinical settings. The host sepsis environment, including the pathogens and the host immune response (cells and soluble factors), may play a crucial role in modifying the therapeutic ePicacy of infused MSCs. Here, we aim to investigate the interplay between MSCs and the host microenvironment in three disease contexts: 1) severe SARS-CoV-2 infection (COVID-19), 2) acute lung injury (ALI) induced by H1N1 influenza A virus (IAV) infection and 3) human sepsis caused by heterogenous organisms. First, we hypothesized that, in severe COVID-19, priming MSCs with a viral mimic would improve their abilities to rebalance the dysregulated immune responses. Transcriptome analysis of Poly(I:C)-primed MSCs (pIC-MSCs) revealed upregulation of pathways involved in antiviral and immunomodulatory responses. Together with increased expression of antiviral proteins, these changes translated into greater MSC ePector functions in regulating monocytes and granulocytes, while enhancing their ability to block SARS-CoV-2 pseudovirus entry into epithelial cells. Importantly, the addition of pIC-MSCs to COVID-19 patient whole blood significantly reduced inflammatory neutrophil populations, increased the proportion of M2 monocytes and enhanced their phagocytic function.
In the second study, deep immune profiling was performed on airway and circulating immune cells to examine the ePect of immunomodulation and therapeutic outcomes of MSCs therapy in mice with H1N1-induced ALI. Immune cell populations and phenotypic shifts were mapped in whole blood by mass cytometry, showing altered immune responses in animals receiving MSCs vs vehicle treatment. Compared to sham animals, IAV infection induced a significant increase in BAL total cell counts. MSC administration significantly decreased BAL total cell counts and altered immune infiltrations in IAV-infected mice.
Phenotypic immune cell profiling of blood and BAL revealed a significant increase in the monocyte population with M2 phenotype in MSCs-treated animals. However, MSCs treatment did not improve survival of infected mice or reduce viral titres in the lungs of infected mice. Further investigation revealed that MSCs were highly susceptible to H1N1 IAV infection, leading to increased cell death and potentially reduced their ePicacy. Lastly, ex vivo human and in vivo animal models were utilized to investigate neutrophil and monocyte modulation by MSCs in sepsis. Results showed phenotypic and functional improvements of myeloid cells after coculture with MSCs. Interestingly, transcriptomic and cytokine analyses revealed adaptive reprograming of MSCs in response to the sepsis milieu. This activation of caspase-1 pathway of MSCs was subsequently confirmed as an essential molecular mechanism for the immunomodulatory ePects of MSCs on dysfunctional neutrophils and monocytes. Caspase-1 activation in MSCs enhanced their therapeutic benefits in treating sepsis in vivo.
Overall, my work suggests that the function and ePicacy of MSCs vary depending on types of infection induced by diPerent pathogens and the host immune microenvironment. These findings highlight the importance of understanding the adaptive responses of MSCs to the host environment in order to better tailor MSC-based therapy for severe viral and bacterial infections
Gestational Diabetes Mellitus: An Exploration of Temporal Trends in Maternal and Neonatal Outcomes in Ontario and a Systematic Review of Benefits of Induction
No full textBackground
Gestational diabetes mellitus (GDM) is a common pregnancy complication that has increased substantially worldwide over the past decades. However, there are few data on recent trends in GDM and the associated impact on adverse pregnancy outcomes in the Canadian population. Induction of labour is recommended to improve perinatal outcomes for GDM pregnancies reaching term gestation, however, epidemiological studies have yielded inconsistent results. The three primary objectives of this thesis were to (1) describe recent trends in GDM and their associations with common risk factors; (2) assess temporal trends in the rates of GDM-associated adverse maternal and neonatal outcomes and compare risks of these outcomes by delivery period; (3) assess the impact of induction of labour at term gestation on adverse outcomes among pregnancies with GDM.
Methods
A population-based retrospective cohort study of all pregnant individuals who had a singleton hospital birth (live birth and stillbirth) between April 1, 2012 and March 31, 2020 in Ontario, Canada, was used to address thesis objectives (1) and (2). Decomposition and multivariable regression analyses with generalized estimation equations were used to evaluate temporal trends and associations with changing trends in GDM. A systematic review and meta-analysis of experimental and observational comparative studies was undertaken to address thesis objective (3).
Results
Objective (1)
Among 1 044 258 pregnant individuals, the annual rate of GDM increased significantly between 2012/13 and 2019/20, from 6.1 to 10.4 per 100 deliveries, in Ontario, Canada (adjusted relative risk:1.53, 95% CI [1.50, 1.56]). The increase was greater in those who did not receive pharmaceutical treatments than those who received pharmaceutical treatments, especially in more recent years. Twenty seven percent of the temporal increase in GDM can be attributed to the distributional changes of three risk factors: the increasing prevalence of advanced maternal age at delivery, pre-pregnancy obesity, and Asian race/ethnicity.
Objective (2)
Among adverse maternal outcomes, despite temporally increasing rates of induction in both GDM and non-GDM pregnancies, the magnitude of the association between GDM and induction was similar between 2012/13-2015/16 and 2016/17-2019/20 (1.62, 95% CI [1.60, 1.64] vs 1.60, 95% CI [1.59, 1.62]). The adjusted relative risk of pregnancy-induced hypertension for GDM compared to non-GDM attenuated from 1.45 (95% CI 1.41, 1.49) in 2012/13-2015/16 to 1.29 (95% CI 2.25, 2.32) in 2016/17-2019/20. The association between GDM and CS (1.10, 95% CI [1.08, 1.12] vs 1.07, 95% CI [1.05, 1.08]), and assisted vaginal delivery (0.96, 95% CI [0.92, 1.00] vs 0.94, 95% CI [0.90, 0.98]) remained similar between the two time periods. For overall maternal morbidity and mortality, the adjusted relative risk was close to 1 and not statistically significant in either time period (0.93, 95% CI [0.78, 1.08] vs 1.09, 95% CI [0.97, 1.20]).
Among adverse neonatal outcomes, although the rate of overall severe neonatal morbidity and mortality (SNM) increased over time among both pregnancies with and without GDM, the positive association between GDM and SNM was stable between the time periods (1.50, 95% CI [1.48, 1.52] vs 1.54, 95% CI [1.52, 1.56]). The magnitude of the association of GDM with LGA and macrosomia was lower in 2016/17-2019/20 than 2012/13-2015/16 (1.42, 95% CI [1.39, 1.45] vs 1.31, 95% CI [1.28, 1.34] for LGA; 0.98, 95% CI [0.95, 1.02] vs 0.86, 95% CI [0.83, 0.90] for macrosomia). There were no significant changes in the associations of GDM with other neonatal outcomes.
Objective (3)
The systematic review identified 11 experimental and observational comparative studies. Random-effects meta-analysis demonstrated that compared to expectant management, induction at term gestation was associated with a lower pooled odds of macrosomia (randomized controlled trials: 0.49, 95% CI [0.30, 0.81]; I2=0%; observational studies: 0.64, 95% CI [0.54, 0.77]; I2=0%); and severe perineal lacerations (observational studies: 0.59, 95% CI [0.39, 0.88]; I2=0%) among GDM pregnancies. There were no significant differences in odds of CS or other adverse perinatal outcomes between the two groups.
Conclusion
There has been a substantial increase in the GDM rates in Ontario from 2012/13 to 2019/20, with twenty seven percent of this increase attributed to the higher prevalence of advanced maternal age at delivery, pre-pregnancy obesity, and Asian ethnicity. On the other hand, despite the increasing rate of GDM, no significant increase in the risk of GDM-associated adverse maternal and neonatal outcomes was observed, except for postpartum hemorrhage with interventions. For pregnant individuals with GDM, induction at term gestation could lower the risk of macrosomia and severe perineal lacerations compared to expectant management. Future well-designed clinical trials with large sample sizes are needed for better screening, diagnosis, and management of GDM
AddShare+: An Efficient Additive Secret Sharing Approach for Private Federated Learning
No full textFederated Learning (FL) has emerged as a transformative approach for collaboratively training Machine Learning (ML) models without the need to share raw data. Despite its benefits, FL is susceptible to significant privacy threats, such as inference attacks that allow adversaries to reconstruct training data from shared model updates and eavesdropping attacks where malicious actors intercept communications to access sensitive information. This thesis addresses these challenges by introducing the AddShare and AddShare+ frameworks, which leverage additive secret sharing and advanced encryption techniques as means of protection during the FL process.
The AddShare framework transforms model updates into multiple shares distributed among clients, rendering it impossible for any single client or adversary to reconstruct the original model. AddShare offers three primary advantages: enhanced privacy through additive secret sharing, secure data transmission using Rivest–Shamir–Adleman (RSA) encryption, and improved efficiency via client grouping. By organizing clients into groups and distributing shares within these groups, AddShare significantly reduces communication overhead while maintaining robust privacy guarantees.
Building on these foundations, AddShare+ introduces two key enhancements. First, it employs selective weight sharing, where only a subset of model weights are shared, reducing computational and communication costs. AddShare+ utilizes a magnitude-based weight selection strategy, focusing on the most impactful weights to optimize the balance between model performance and privacy. Second, it replaces RSA with the more efficient Elliptic Curve Integrated Encryption Scheme (ECIES), enabling faster and more secure share distribution. These frameworks ensure strong privacy preservation while maintaining model accuracy, making them suitable for sensitive applications such as healthcare, finance, and the Internet of Things (IoT).
Comprehensive experimental evaluations demonstrate the efficacy of these frameworks. For instance, AddShare+ achieves comparable model accuracy to traditional FL methods on the Fashion - Modified National Institute of Standards and Technology (F-MNIST) dataset, maintaining a Server-side Accuracy (SA) of 78.9%, compared to 79.3% with Federated Averaging (FedAvg), while reducing Federated Learning Time (FLT) by 13% relative to AddShare.
The thesis concludes with a case study on crop yield prediction, showcasing the practical application of AddShare+ in a smart farming context. While aggregated farming data holds immense potential for improving agricultural productivity and sustainability through ML and FL, farmers are often reluctant to share their operational data due to concerns about its impact on loan negotiations, insurance rates, land valuations, and competitive advantages in local markets. This real-world scenario demonstrates how the proposed framework can be applied to sensitive agricultural data, balancing the need for collaborative learning with data privacy concerns. These frameworks significantly enhance the privacy and security of FL systems against inference and eavesdropping attacks, promoting their broader adoption in various critical applications. Future research directions include further optimization of these frameworks and exploration of additional privacy-preserving techniques to enhance the security and efficiency of decentralized ML systems
Interpretation and Prediction of the Hydromechanical Behavior of Unsaturated Soils from Tropical Regions
No full textTropical soils are widespread in many countries and regions of the world, including Brazil, large portions of the African continent, and Asia. These soils are often found in an unsaturated state and their hydromechanical behavior is predominantly influenced by their structure which is often characterized by intra and inter-aggregate pores.
The intense weathering associated with high temperatures and relatively high annual rainfall results in the formation of aggregations associated with the cementation of clay particles. The top few meters of soil depth profile in tropical regions constitute a highly weathered layer (typically, lateritic soil) sitting on top of a less weathered layer with traces of the parent rock (normally, saprolitic soil). These lateritic soils attain aggregations that impart a dual-porosity resulting in a different pore size characteristics that contribute to bimodal soil-water characteristic curve (SWCC) behavior. Also, they attain different characteristics over their long formation period in comparison to conventional soils; due to this reason, traditional classification systems are not suitable.
Despite advances in recent decades, there are limited studies with respect to interpretation and prediction of the hydro-mechanical behavior of bimodal lateritic soils. For this reason, there is a need for rational tools for use in conventional geotechnical engineering practice for the design of various geo-infrastructures extending the principles of unsaturated soil mechanics that are simple, especially in the context of climate change in which unprecedent changes in suction arise.
Most existing constitutive models describing the behavior of unsaturated soils were developed for soils with a single pore-size family. As a result, these models have limitations or are inadequate for lateritic soils that are typically bimodal in nature. Therefore, the two major objectives of this thesis are: (i) to model the bimodal SWCC, and (ii) to model the shear strength with respect to suction for these dual porosity soils. In this context, the modeling effort includes both advances in the understanding of soil behavior and the development of prediction models.
The SWCC behavior of bimodal lateritic soils was investigated by considering physicochemical aspects and their relation to the soil structure. The analyses were performed relying on a database of 27 soils divided into 25 and 15 datasets corresponding to undisturbed and remolded conditions, respectively. A new framework was proposed to estimate the bimodal SWCC for these soils based on a function describing the relation between particle and pore-size. The framework uses basic soil information and relates the macro and microstructure to the aggregated and disaggregated grain-size distribution (GSD) curves, respectively. Adsorption was indirectly considered by incorporating simple soil properties such as the liquid limit in the calibration of the function parameters. The coefficient of uniformity and the degree of aggregation were found to be associated with the desaturation zones of the macro and micropores.
A simplified prediction model for the SWCC based on correlations and nonlinear regression approaches was also proposed. The regression analyses indicated that the level of aggregation and the liquid limit are the most relevant factors affecting microstructure whereas the macrostructure is strongly governed by parameters from the aggregated GSD curve. The performance of both models evaluated through R² is reasonably good, with values consistently exceeding 0.80.
The shear strength model involved the analysis of the evolution of soil structure during shearing and its impact on the relationship between matric suction, net normal stress, and shear strength. Such analyses included the interpretation of the unusual behavior of some bimodal lateritic soils, where the contribution of suction to shear strength is greater than that of net normal stress (i.e., ϕb > ϕ'). The model assumes that only suctions within the micropores zone cause relevant structural changes during shearing that can decrease ϕb whereas ϕ' varies in all suction ranges. At high suctions, the contribution of suction to the shear strength becomes constant. These assumptions constitute the basis for a new shear strength prediction model that is easy to implement, requires only GSD and SWCC information, and offering superior results (R² > 0.95) in comparison to 14 unimodal and 2 bimodal models available in the literature. The increase in the effective friction angle, ϕ' with respect to matric suction during shearing seems to be a function of the level of aggregation; however, the values of the friction angle associated with matric suction, ϕb, greater than ϕ' are believed to occur due to the emergence of a contractive shear band in specimens with suctions higher than the first air-entry value in drained conditions. The relationship between apparent cohesion and matric suction appears to exhibit a linear relationship extending from saturation to the end of the macropores region, and a non-linear relationship starting from the preceding limit to the midpoint of the micropores transition region.
Although the purpose of the models presented in this thesis is not to replace direct determinations using laboratory tests, they can be considered a step forward in the implementation of unsaturated soil mechanics for lateritic soils because they offer reasonable estimates of the SWCC and the unsaturated shear strength. This information is essential in developing preliminary geotechnical designs, such as those related to slope stability, retaining walls, bearing capacity of formation layers of roads and railroads, bearing capacity in foundations, cover and capillary barrier systems, and mining waste disposal systems
Speech Production Effects on Word Learning and Memorization in Children
No full textIn this dissertation, I explore the impact of speech production on memory and learning in children, by investigating the Production Effect and the Reverse Production Effect. In the Production Effect, a memory advantage is observed for items that are produced aloud compared to items studied under other actions, such as reading silently or listening. Conversely, the Reverse Production Effect is a memory advantage for items studied under listening conditions, compared to producing aloud. Despite extensive research showing the Production Effect with adults, its manifestation and underlying mechanisms in children remains largely unexplored. This dissertation focuses on children, covering an age range from 2-to-6-years-old across all experiments combined. Different methodologies were used to study memory and learning in monolingual English-speaking children, specifically free recall tasks, identification tasks, and recognition tasks using the Visual World Paradigm with eye-tracking.
This dissertation presents four articles that investigate different aspects of the Production Effect and Reverse Production Effect, with a larger goal of examining how children's developing cognitive and linguistic abilities may influence the effect of speech production on memory and learning. The first article used a developmental approach to investigate whether the Production Effect changes across development, using a free recall task with real words. While older children (4-to-6-years-old) were able to benefit from speech production, younger children (2-to-3-years-old) showed a Reverse Production Effect, suggesting that the developmental and linguistic stage of the learner interacts with the Production Effect. The second study investigated whether the Production Effect found in children (5-to-6-years-old) remains after a 1-week-delay. Results show that for children who did show the Production Effect initially, this effect was not present when tested on the same words after 1 week, suggesting that the memory advantage is restricted in time. The third article investigated whether the Reverse Production Effect found in children (5-and-6-year-olds) is specific to speech production, or whether other actions would also disrupt learning. All actions patterned similarly, showing a Reverse Production Effect, indicating that this is not solely triggered by speech production, but by performing any action while learning. The fourth article investigated whether a more familiar context (storybook reading) and multiple exposures to the novel items would increase the chance of the Production Effect emerging in a word learning task with children aged 4-and-6-years. Results patterned with previous Reverse Production Effect findings, with further effects of children's age and the number of exposures during training.
These findings deepen our understanding of how speech production impacts memory encoding and retrieval across different stages of development. By examining the Production Effect and Reverse Production Effect in children, this dissertation contributes to psycholinguistic theories of language acquisition and memory, offering valuable insights into learning strategies from early childhood. Results from this dissertation show that production and perception influence memory in distinct ways. Practical implications include potential applications in educational settings
A stress-dependent TDP-43 SUMOylation program preserves neuronal function
No full textAbstract Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Dementia (FTD) are overwhelmingly linked to TDP-43 dysfunction. Mutations in TDP-43 are rare, indicating that the progressive accumulation of exogenous factors – such as cellular stressors – converge on TDP-43 to play a key role in disease pathogenesis. Post translational modifications such as SUMOylation play essential roles in response to such exogenous stressors. We therefore set out to understand how SUMOylation may regulate TDP-43 in health and disease. We find that TDP-43 is regulated dynamically via SUMOylation in response to cellular stressors. When this process is blocked in vivo, we note age-dependent TDP-43 pathology and sex-specific behavioral deficits linking TDP-43 SUMOylation with aging and disease. We further find that SUMOylation is correlated with human aging and disease states. Collectively, this work presents TDP-43 SUMOylation as an early physiological response to cellular stress, disruption of which may confer a risk for TDP-43 proteinopathy
Cryptographic Applications of the Quantum No-Cloning Principle for Messages, Programs, and Advice
No full textThe no-cloning principle tells us that it is impossible to perfectly replicate the state of an arbitrary quantum system. From its early beginnings with Wiesner, Bennett, and Brassard, quantum cryptography has often relied, either implicitly or explicitly, on this principle to achieve security properties beyond the reach of classical cryptography. We examine in this thesis various such cryptographic applications of quantum uncloneability from two broad perspectives: uncloneability for messages and uncloneability for functionalities.
Starting with uncloneability for messages, we study relations between tamper-evident encryption schemes - symmetric-key schemes for classical messages producing quantum encodings which allows the honest recipient to detect meaningful eavesdropping - and various other security notions. We show that tamper evidence implies encryption, that it can be used to construct revocation schemes (and vice-versa), and we formalize Gottesman's construction of quantum money from schemes satisfying this notion.
Moving on to uncloneability for functionalities, we begin by detailing our contributions to the study of quantum copy-protection and secure software leasing. These are two similar but distinct notions of uncloneability for families of functionalities. We show how to instantiate a scheme for point functions which is secure against honest-malicious adversaries in the quantum copy-protection setting and show that this is, generically, a sufficient condition to construct a secure software leasing scheme for the same functions.
We then complete this thesis by reporting work in which we initiate the formal study of uncloneable quantum complexity theory. We accomplish this by defining a novel complexity class denoted neglQP/upoly - the class of promise problems and languages which can be solved with bounded error by efficiently describable quantum circuits with the help of a quantum advice state exhibiting a type of functional uncloneability - and we show, unconditionally, that this class is non-empty. This relates to uncloneable functionalities as we also formalize how uncloneable advice can be seen as a form of quantum copy-protection for specific functions
Critical Consciousness in Sport Scale (CCSS): development and initial psychometrics properties
No full textAbstract Background Critical consciousness refers to the ability to recognize and analyze the oppressive forces in society and to act against them. Researchers have emphasized that sport has the potential to help athletes develop their critical consciousness regarding the forces that shape society, as well as the systems of privilege and deprivation that influence access to sport. Objective The present study describes the development of the critical consciousness in sport scale, offering initial validity evidence based on test content, internal structure, relations with other variables, and reliability. Methods An initial set of 50 items was develop and reviewed by expert judges who assessed the clarity, practical relevance, and theoretical adequacy of the items using the content validity coefficient (CVC). To investigate the internal structure of the CCSS, factor retention methods such as parallel analysis and exploratory graph analysis were employed, followed by an exploratory factor analysis (EFA), along with an assessment of the internal consistency of the factors. Validity evidence based on relations with other variables was estimated using Pearson correlations. The sample was comprised of 263 Brazilian psychology and physical education students (mean age: 26.95 ± 9.69; 70.02% women). Results Factor retention methods that included parallel analysis, exploratory graph analysis, and a categorical exploratory factor analysis demonstrated a three-dimensional structure comprised of 36 items, as theoretically hypothesized, with desirable internal consistency indices (ω = 0.868, 0.906, and 0.924, respectively). A brief version of the instrument is also presented, which adequately reproduced the psychometric properties of the initial version. Correlations with measurements of social justice and anti-racism efficacy suggested validity evidence based on relations with other variables. Conclusion The results suggest that the instrument is an appropriate measure of critical consciousness in sport. It is recommended that future efforts focus on estimating further validity evidence and reliability of the CCSS in diverse samples of athletes from different competitive levels, sports coaches, fans, and other key figures within the sports context
Exploring the intersection of adverse childhood experiences and Attention-Deficit/Hyperactivity Disorder symptoms in adulthood: Differential vulnerability and resilience factors
No full textAttention-Deficit/Hyperactivity Disorder (ADHD) is a prevalent neurodevelopmental disorder that persists into adulthood. Studies have shown that Adverse Childhood Experiences (ACEs) are associated with ADHD symptoms and emotional dysregulation. Additionally, research suggests that resilience factors may play a protective role on the impact of ACEs on ADHD symptoms and emotional dysregulation. This study investigates the relationship between ACEs with ADHD symptom severity and co-occurring Emotion Regulation (ER) in adults, and examines how resilience may moderate the association between ACEs and ADHD symptoms as well as ACEs and ER in adult populations. Three hundred and six participants between the ages of 18 to 55 years with diagnosed or high self-reported ADHD symptoms completed surveys measuring ADHD symptoms, ER, ACES, and resilience. Correlational analyses examined relationships between ADHD symptoms, ACES, ER, and resilience. Significant results showed a negative association between ACES and ER, ADHD symptoms and ER, and ACEs and resilience, along with a positive relationship between resilience and ER. Additionally, exploratory analyses demonstrated that ACEs were associated with the Predominantly Inattentive Presentation of ADHD, and that resilience specific to family cohesion and social resources were potential moderators. These findings highlight the complex interplay between ACEs, resilience, and ADHD symptoms