DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
The neurodevelopmental gradient hypothesis from an executive function perspective
No full textThis paper explores how executive functions and their functional connectivity (FC)
representations are related to the neurodevelopmental gradient hypothesis by Owen and
O’Donovan (2017). Data from 231 participants (aged 21-50) in the UCLA Consortium for
Neuropsychiatric Phenomics LA5c Study (CNP) was used, comprising 113 healthy participants
(HC), 36 with attention-deficit/hyperactivity disorder (ADHD), 43 with bipolar disorder (BD),
and 39 with schizophrenia (SCZ). ANCOV A was used to compare task performance, controlling
for age and gender. FC analyses were done using network-based statistics (NBS) on resting-state
and task data, with post-hoc Pearson’s correlation tests on significant network strengths.
Significant differences in task performance were found in one measure of the color trails task,
spatial working memory task, stop signal reaction time, and task switching accuracy, while no
significant differences were found for the color trails interference index or continuous
performance task. Planned comparisons suggested that executive dysfunction was most severe in
SCZ, with performance in ADHD and BD being worse than in HC but ambiguous in relation to
one another. No significant results were found for resting-state FC analysis, while significant
findings were present only for the SCZ group in two tasks. Post-hoc correlational analyses
suggested significant but weak correlations between network strength and task performance. The
results suggest a divergence from Owen and O’Donovan’s (2017) hypothesis, though limitations
of the sample should be accounted for. Future research should seek to understand transdiagnostic
dimensions of mental illnesses and understand how genetics translate to psychological function
to further knowledge.Bachelor's degre
Web application for promoting lifelong learning
No full textLifelong learning has been a recognized concept since its formal adoption by UNESCO in the 1960s, yet its implementation has often been inconsistent or overlooked. Recent national initiatives, such as Singapore’s SkillsFuture and efforts by institutions like Nanyang Technological University (NTU), have revitalized interest in continuous education. In this context, an intelligent online platform presents a timely and relevant solution to promote lifelong learning and support career development.
The proposed platform is architected as an intelligent, user-centric ecosystem that redefines how individuals engage with lifelong learning. Designed with a deep sensitivity to usability, the system delivers a highly personalized experience through dynamic career recommendations, informed by nuanced interpretation of user-provided resumes and guided inputs. Its core engine synthesizes user data to surface tailored career pathways that are both aspirational and attainable, aligned with each individual’s evolving capabilities.
Beyond static suggestions, the platform supports immersive career exploration, allowing users to interactively navigate potential trajectories based on latent skill affinities. Once a target pathway is identified, the platform initiates a high-resolution skill gap assessment, dynamically mapping learning objectives to curated educational content. A responsive dashboard overlays this journey with real-time progress visualization, encouraging iterative growth and self-reflection. The result is a seamless and adaptive experience — one that not only empowers users with insight, but also motivates sustained engagement through clarity, structure, and actionable next steps.Bachelor's degre
Induction of cellular senescence in RPE1 cells through DNA damage response pathway
No full textAge-related macular degeneration (AMD) is closely linked to the accumulation of cellular damage and senescence in retinal pigment epithelium (RPE) cells. To model senescence in vitro, we treated RPE1 cells with 30uM of etoposide, a DNA topoisomerase II inhibitor known to induce DNA double-strand breaks (DSBs), triggering a senescence program. Etoposide-treated cells exhibited hallmark features of senescence, including increased cell size, enhanced SA-β-gal activity, elevated γH2AX foci, reduced population doubling level (PDL), and robust upregulation of CDKN1A transcripts. Beyond nuclear markers, we characterized subcellular perturbations associated with senescence. Mitochondria in etoposide-treated cells displayed elongated morphology suggestive of dysfunction, and lysosomes appeared enlarged and more abundant, consistent with compromised autophagic clearance. Endoplasmic reticulum (ER) morphology was notably altered, with increased distention and circular ER-derived bodies, implicating ER stress. Furthermore, etoposide-induced senescence led to transcriptional downregulation of adaptive unfolded protein response (UPR) branches, namely IRE1 and ATF6, while PERK activity remained unchanged. These impairments were reinforced by diminished XBP1s induction upon tunicamycin induction. Collectively, our findings demonstrate that etoposide-induced DNA damage is sufficient to elicit a robust senescence phenotype in RPE1 cells and uncover multi-layered organelle dysfunction involving mitochondria, lysosomes, ER architecture, and UPR signalling, highlighting mechanistic intersections relevant to AMD pathogenesis.Bachelor's degre
Research on indoor sound localization and event detection
No full textThe purpose of this research is to investigate indoor sound source localization and sound event classification technologies. The study is divided into two sections, the first of which addresses the issue of localizing indoor sound sources. The second section focusses on indoor sound event classification using the SINS dataset and builds models based on VGGish and ResNet network structures, respectively. It suggests a generalized cross-correlation (GCC) method to compute the time delay difference between microphone arrays and combines the space search algorithm with the least squares iteration method to estimate the direction of sound. This dissertation presents a range of data augmentation and feature augmentation strategies, such as head-related transfer function (HRTF), harmonic-transient-noise separation (HPSS), and mix-up and masking methods, to increase the model's classification accuracy. According to the experimental results, the model's robustness in complex situations and macro accuracy have both greatly increased when the augmentation technique was implemented.Master's degre
Mechanisms of temperature control of singlet fission in an optical cavity
No full textWe investigate the mechanisms of temperature control in conical-intersection-mediated singlet fission (SF) within optical cavities. Using the multiple Davydov D2 Ansatz combined with the thermo-field dynamics formalism, we model the quantum dynamics of a rubrene dimer coupled to an optical cavity at finite temperatures. The work explores the influence of temperature, cavity-matter coupling strength, photon frequency, and cavity loss on the triplet–triplet population dynamics. Results reveal that temperature enhances SF efficiency via thermal activation of coupling modes and assists in overcoming potential barriers between singlet and triplet states. It is found that strong photon-matter coupling and high photon frequencies also promote SF under conditions of resonance with excited vibronic states, while cavity losses and increased photon numbers can inhibit the process. Increased average photon numbers suppress SF as the polaritonic conical intersections shift away from the Franck–Condon region, although a photon-assisted SF effect is revealed for specific values of the average photon number at low temperatures. The study provides insights into the temperature control mechanisms of SF in optical cavities, offering potential directions for designing functional optical cavities to enhance SF efficiency, with implications for organic photovoltaics and other energy transfer technologies.Ministry of Education (MOE)Submitted/Accepted versionOne of us (M.F.G.) thanks the support from the National Natural Science Foundation of China (Grant No. 22373028). The authors also gratefully acknowledge the support of the Singapore Ministry of Education Academic Research Fund (Grant No. RG2/24)
Collision prediction and validation for a 3D upper limb end-effector robot in reachable workspace
No full textThis dissertation focuses on the development and validation of a collision detection
system for assistive robotic arm extenders, designed to enhance the safety of upper
limb rehabilitation robots. The primary objective of this research is to ensure safe
human-robot interaction by accurately detecting potential collisions during
rehabilitation exercises, thereby preventing harm to the user.
The study proposes an algorithm for collision prediction. A kinematic model is
formulated for both the human and robotic arms, utilizing the Denavit-Hartenberg
(D-H) method to derive the forward and inverse kinematics of each system. To
predict potential collision regions, two advanced methodologies are employed: the
Monte Carlo method and the Convex Hull Set method. These methods are
instrumental in defining the reachable workspace of both the human arm and the
robotic system, thus laying the foundation for effective collision detection.
The collision detection algorithm is rigorously validated through both theoretical
and experimental approaches. Theoretical validation involves simulation studies of
collision and non-collision scenarios, providing a preliminary evaluation of the
algorithm's applicability. Experimental validation is performed using motion
capture (Mocap) data and real-world collision tests, facilitating a direct comparison
between theoretical predictions and empirical results. The comparative analysis
demonstrates the algorithm's capability to accurately detect potential collisions and
underscores its potential for real-time application in rehabilitation
environments.The dissertation also outlines future improvements, including
optimizing the detection algorithm for dynamic interactions and developing
real-time control strategies to reduce collision risks.Master's degre
Degradation of persistent micropollutant using ZIF-67 MXene as catalyst
No full textWith increasing consumerism, there has been more investments in the commercial
sector, especially in healthcare and cosmetics. Organic compounds are widely found in the
healthcare and cosmetics sector such as in pharmaceuticals, cosmetics, and personal care
products. Due to the increased consumption of these products, organic compounds have been
increasingly prevalent in water systems. Although the manufacturing and consumption of
organic compounds benefit the society, they pose significant environmental and health risks
even at low concentrations. Although conventional treatment methods have been investigated
and implemented in an attempt to remove these pollutants in water systems, it faces limitations
in removing these organic compounds due to the wide variety available and complex chemical
properties. Hence, advanced solutions such as Peroxymonosulfate-based Advanced Oxidation
Processes (PMS-AOP) have been an growing interest. However, some challenges are faced by
this technology such as lower mass transfer efficiency and agglomeration. Thus, novel solutions
are needed to overcome and reduce these limitations.
This study focuses on the exploration and fabrication of the ZIF-67 MXene catalyst, a
metal organic framework combined with 2D material of MXene as an emerging nanocomposite
to enhance their properties and become a potential solution to degrading organic compounds.
The fabrication procedure crafted from studies and research done were modified and improved
throughout the project to improve the morphology and performance of the ZIF-67 MXene
catalyst.
After fabricating the ZIF-67 MXene catalyst, its performance was analysed through
degrading three different pollutants with different chemical bond strength and properties, Acid
Orange 7 (AO7) , Enrofloxacin and PFOA. The catalyst managed to increase the degradation
efficiency of PMS which was used as an oxidant by 85% but only 22.5% for enrofloxacin
degradation. However, the results obtained for PFOA showed that the catalyst was not as
effective in enhancing degradation performance due to the difficulty in degradation of PFOA
The catalyst samples also undergo a series of characterization tests such as FESEM,
XRD and FTIR which shows the morphology and chemical properties of the samples to ensure
that the catalyst was successfully fabricated and also aids in improving fabrication proceduresBachelor's degre
Development of modular sensorised tools for assessing infant executive function development
No full textThis report outlines the progress of a final year project aimed at designing and developing modular sensorised tools to assess the executive function (EF) development in infants, focusing on working memory. The system will record infant response times, number of attempts, and other relevant metrics to provide a quantitative evaluation of working memory. This report details the project's objectives, methodology, instrumented toy design, preliminary results, and future steps.Bachelor's degre
Recent fluorination strategies in solid electrolytes for high-voltage solid-state lithium-ion batteries
No full textHigh-voltage solid-state lithium-ion batteries (SSLIBs) have attracted considerable research attention in recent years due to their high-energy-density and superior safety characteristics. However, the integration of high-voltage cathodes with solid electrolytes (SEs) presents multiple challenges, including the formation of high-impedance layers from spontaneous chemical reactions, electrochemical instability, insufficient interfacial contact, and lattice expansion. These issues significantly impair battery performance and potentially lead to battery failure, thus impeding the commercialization of high-voltage SSLIBs. The incorporation of fluorides, known for their robust bond strength and high free energy of formation, has emerged as an effective strategy to address these challenges. Fluorinated electrolytes and electrode/electrolyte interfaces have been demonstrated to significantly influence the reaction reversibility/kinetics, safety, and stability of rechargeable batteries, particularly under high voltage. This review summarizes recent advancements in fluorination treatment for high-voltage SEs, focusing on solid polymer electrolytes (SPEs), inorganic solid electrolytes (ISEs), and composite solid electrolytes (CSEs), along with the performance enhancements these strategies afford. This review aims to provide a comprehensive understanding of the structure–property relationships, the characteristics of fluorinated interfaces, and the application of fluorinated SEs in high-voltage SSLIBs. Further, the impacts of residual moisture and the challenges of fluorinated SEs are discussed. Finally, the review explores potential future directions for the development of fluorinated SSLIBs.Agency for Science, Technology and Research (A*STAR)Nanyang Technological UniversityNational Research Foundation (NRF)Submitted/Accepted versionThis study was financially supported by the A*STAR MTC Programmatic Project (No. M23L9b0052), the Indonesia-NTU Singapore Institute of Research for Sustainability and Innovation (INSPIRASI) (No. 6635/E3/KL.02.02/2023), the Singapore NRF Singapore-China Flagship Program (No. 023740-00001), the National Natural Science Foundation of China (Nos. 11975043 and 11475300), and the China Scholarship Council (No. 202306460087)
Development of a Raspberry Pi based cortisol detector
No full textCortisol concentration is a key biomarker for assessing stress response and
immune function, making it an important indicator of anxiety levels. However,
traditional detection methods, such as blood tests, are often costly and complex.
This study presents a low-cost and user-friendly cortisol detector based on the
Raspberry Pi platform.
The device employs a Raspberry Pi 4B as the central processing unit, integrating
a camera module, LED system, and LCD display. It captures images of
cortisol test strips based on the lateral flow immunoassay (LFIA) method, offering
a rapid and noninvasive detection approach. Python scripts control LED
illumination and image acquisition, while OpenCV-based image processing extracts
intensity data from control (C) and test (T) lines to quantify cortisol
concentration. Results are displayed in real time on the LCD screen.
The system’s affordability and ease of use make it suitable for clinical and
home applications. Future improvements may enhance automation and enable
cloud-based data storage, further promoting stress monitoring and health assessment.Master's degre