Dylan Amiri

Integrating artificial intelligence (AI) in healthcare has sparked innovation but exposed vulnerabilities in regulatory oversight. Unregulated “shadow” AI systems, operating outside formal frameworks, pose risks such as algorithmic drift, bias, and disparities. The Comprehensive Algorithmic Oversight and Stewardship (CAOS) Framework addresses these challenges, combining risk assessments, data protection, and equity-focused methodologies to ensure responsible AI implementation. This framework offers a solution to bridge oversight gaps while supporting responsible healthcare innovation. CAOS functions as both a normative governance model and a practical system design, offering a scalable framework for ethical oversight, policy development, and operational implementation of AI systems in healthcare.

Ophthalmic diseases such as glaucoma, age-related macular degeneration (ARMD), and optic neuritis involve complex molecular and cellular disruptions that challenge current diagnostic and therapeutic approaches. Advanced artificial intelligence (AI) and machine learning (ML) models offer a novel lens to analyze these diseases by integrating diverse datasets, identifying patterns, and enabling precision medicine strategies. Over the past decade, applications of AI in ophthalmology have expanded from imaging-based diagnostics to molecular-level modeling, bridging critical gaps in understanding disease mechanisms. This paper systematically reviews the application of AI-driven methods, including reinforcement learning (RL), graph neural networks (GNNs), Bayesian inference, and generative adversarial networks (GANs), in the context of these ophthalmic conditions. RL models simulate transcription factor dynamics in hypoxic or inflammatory environments, offering insights into disrupted molecular pathways. GNNs map intricate molecular networks within affected tissues, identifying key inflammatory or degenerative drivers. Bayesian inference provides probabilistic models for predicting disease progression and response to therapies, while GANs generate synthetic datasets to explore therapeutic interventions. By contextualizing these AI tools within the broader framework of ophthalmic disease management, this review highlights their potential to transform diagnostic precision and therapeutic outcomes. Ultimately, this work underscores the need for continued interdisciplinary collaboration to harness AI’s potential in advancing the field of ophthalmology and improving patient care.

Advancements in neuroimaging, particularly diffusion magnetic resonance imaging (MRI) techniques and molecular imaging with positron emission tomography (PET), have significantly enhanced the early detection of biomarkers in neurodegenerative and neuro-ophthalmic disorders. These include Alzheimer’s disease, Parkinson’s disease, multiple sclerosis, neuromyelitis optica, and myelin oligodendrocyte glycoprotein antibody disease. This review highlights the transformative role of advanced diffusion MRI techniques—Neurite Orientation Dispersion and Density Imaging and Diffusion Kurtosis Imaging—in identifying subtle microstructural changes in the brain and visual pathways that precede clinical symptoms. When integrated with artificial intelligence (AI) algorithms, these techniques achieve unprecedented diagnostic precision, facilitating early detection of neurodegeneration and inflammation. Additionally, next-generation PET tracers targeting misfolded proteins, such as tau and alpha-synuclein, along with inflammatory markers, enhance the visualization and quantification of pathological processes in vivo. Deep learning models, including convolutional neural networks and multimodal transformers, further improve diagnostic accuracy by integrating multimodal imaging data and predicting disease progression. Despite challenges such as technical variability, data privacy concerns, and regulatory barriers, the potential of AI-enhanced neuroimaging to revolutionize early diagnosis and personalized treatment in neurodegenerative and neuro-ophthalmic disorders is immense. This review underscores the importance of ongoing efforts to validate, standardize, and implement these technologies to maximize their clinical impact.

Advancements in neuroimaging, particularly diffusion MRI techniques and molecular imaging via PET, have significantly improved our ability to detect early biomarkers of both neurodegenerative and neuro-ophthalmic disorders, such as Alzheimer's, Parkinson's, multiple sclerosis (MS), neuromyelitis optica (NMO), and myelin oligodendrocyte glycoprotein (MOG) antibody disease. Despite these breakthroughs, the integration of AI-driven models has become crucial in harnessing the full potential of these technologies for clinical decision-making. This paper explores the role of advanced diffusion MRI techniques—specifically Neurite Orientation Dispersion and Density Imaging (NODDI) and Diffusion Kurtosis Imaging (DKI)—in revealing microstructural changes in brain and visual pathway tissues that precede clinical symptoms. These techniques, when paired with AI algorithms, enhance diagnostic precision, enabling the detection of early-stage degeneration and inflammatory processes with unprecedented accuracy. Additionally, the emergence of next-generation PET tracers targeting misfolded proteins, such as tau and alpha-synuclein, and inflammatory markers, further augments our ability to visualize and quantify pathological processes in vivo. The integration of AI, particularly deep learning models like convolutional neural networks (CNNs) and multimodal transformers, has demonstrated remarkable success in improving diagnostic accuracy, combining data from multiple imaging modalities, and providing predictive insights into disease progression. This review also examines the challenges associated with integrating AI into clinical practice, including technical variability, data privacy concerns, and regulatory hurdles. While these technologies promise to revolutionize early diagnosis and personalized treatment approaches, significant efforts in model validation, standardization, and clinical implementation remain crucial. The ongoing development of AI-enhanced neuroimaging holds great potential for advancing the precision of diagnosis and the effectiveness of therapeutic interventions in both neurodegenerative and neuro-ophthalmic disorders.

This review examines recent advancements and persistent challenges in corticospinal tract (CST) visualization, highlighting its critical role in motor function assessment and neurosurgical planning. Despite improvements in imaging techniques, CST visualization continues to face issues related to variability in imaging protocols, data processing methods, and visualization standards, which impact consistency and reproducibility across studies. We explore various AI-driven and machine learning (ML) approaches that offer promising solutions, including high-resolution diffusion MRI (dMRI) tractography, interactive 3D models, and virtual reality systems, which collectively enhance CST interpretability by refining spatial resolution and fiber orientation estimation. Additionally, automated visualization techniques using AI facilitate adaptive visualization, making CST representations more consistent and clinically relevant. High-resolution CST visualization holds potential clinical applications in neurology and neurosurgery, particularly for diagnosing and managing stroke, multiple sclerosis (MS), and concussion—disorders where motor pathway integrity is crucial for treatment planning and prognosis. However, a lack of standardization and interoperability among CST visualization tools remains a significant barrier, impeding cross-study comparisons and clinical integration. This review underscores the need for standardized protocols and collaborative efforts within the neuroimaging community to establish best practices, enabling reliable and actionable CST visualization in clinical and research settings. Addressing these challenges may further improve diagnostic accuracy, therapeutic planning, and the overall understanding of motor-related neurological disorders.

Introduction: A common practice in clinical settings is the use of the Epworth Sleepiness Scale (ESS) and apnea-hypopnea index (AHI) to demonstrate the severity of obstructive sleep apnea (OSA). However, several instances were noted where there were discrepancies in the reported severity between Epworth scores and AHI in our patient sample, prompting an investigation into whether OSA severity as demonstrated by AHI or predicted by ESS quantification of sleepiness is primarily responsible for inconsistencies.Methods: Discrepancies were examined between Epworth scores and AHI by categorizing patients into two categories of inconsistency: individuals with either ESS < 10 and AHI ≥ 15 events/h or ESS ≥ 10 and AHI < 15 events/h. The potential influence of sex on these categories was addressed by assessing whether a significant difference was present between mean Epworth scores and AHI values for men and women in the sample. We investigated BMI both by itself as its own respective variable and with respect to the sex of the individuals, along with a consideration into the role of anxiety. Furthermore, we tested anxiety with respect to sex.Results: In the first category of inconsistency the average ESS of 5.27 ± 0.33 suggests a normal level of daytime sleepiness. However, this contrasts with the average AHI of 32.26 ± 1.82 events/h which is indicative of severe OSA. In the second category the average ESS of 14.29 ± 0.47 suggests severe daytime sleepiness, contradicting the average AHI of 9.16 ± 0.44 events/h which only indicates mild OSA. Sex, BMI (both as a variable by itself and with respect to sex), and anxiety (both as a variable by itself and with respect to sex) contributed to observed inconsistencies.Conclusion: The findings of our study substantiate our hypothesis that Epworth scores should be de-emphasized in the assessment of OSA and a greater importance should be placed on measures like AHI. While Epworth scores offer insights into patients’ daytime sleepiness levels and the perceived severity of their OSA, the inconsistencies highlighted in our results when compared to AHI-based OSA severity underscore their potential inaccuracy. Caution is advised when utilizing Epworth scores for evaluating OSA severity in clinical settings.

A required aspect for the classification of a compound as aromatic is it must have a conjugated pi system, which is constituted by a region of overlapping p-orbitals or, in other words, a continuous system of sp2-hybridized atoms. This is usually determined by drawing a valid resonance structure (one with no more than 2 formal charges) where all atoms are sp2-hybridized for the compound of interest, but we propose a new method intuitively suggesting resonance (if the atom of interest is determined to be sp2-hybridized) involving the comparison of inclusive vs. non-inclusive steric number calculation. We define inclusive steric number calculation as the number of sigma bonds plus the number of lone pairs bonded to the atom (both localized and delocalized), and define non-inclusive as the number of sigma bonds plus the number of localized lone pairs only. In the non-inclusive steric number calculation, the delocalized lone pairs count towards the Huckel number of pi electrons in the compound (4n+2 where n is an integer). Once both are calculated for the atom of interest and the rest of the following requirements for aromaticity are evaluated (i.e. compound is planar, cyclic, and has Huckel number of electrons) the version of the compound with the most stable aromatic configuration is selected as the correct configuration. For alkyne-containing compounds, we devised an alternative method for determining sp2-hybridization of atoms involved in an alkyne to ultimately reveal a conjugated pi system in the compound that agrees with our inclusive vs. non-inclusive steric number calculation. This approach involves the idea that the pi electrons of a pi bond within an alkyne that are not involved in the conjugated pi system (due to these p-orbitals being perpendicular relative to the parallel p-orbitals contributing to conjugation) can occupy a single atom in the alkyne, but the specific atom they occupy is unclear. Therefore, we present a “lone pair hybrid” that accounts for both atoms being equally able to receive the pi electrons as a lone pair. These hypothetical lone pairs are localized due to an alkene still bonding the atoms and makes the steric number calculation for both the inclusive and non-inclusive approach 2 sigma bonds + 1 lone pair = 3, sp2, making the atoms sp2-hybridized and thus revealing the conjugated pi system.