Alzheimer’s disease (AD) is an irreversible neurodegenerative disease associated with memory loss, decline in cognitive abilities, and a change in behavior. The most important factor that leads up to Alzheimer’s disease for an individual is an aging brain. As people grow older, brain atrophy, or brain shrinkage, occurs. According to the Alzheimer's Association, in the United States, nearly 7 million people are living with Alzheimer’s, the majority over 65 years of age (Alzheimer’s Association 2025). Inside the brain of an AD patient, amyloid plaques (which look like abnormal clumps located primarily in the amygdala and hippocampus) and tau protein tangles located in neurons (which block the neuron’s transport system) both contribute to Alzheimer’s symptoms. Currently, there is no cure for Alzheimer's disease. Therefore, researchers do focus heavily on early detection and prevention of the disease. In my research proposal, I will discuss public research that supports the argument that artificial intelligence can vastly improve early detection, specifically in terms of neuroimaging and biomarkers. However, Artificial Intelligence is not perfect. There are still flaws and lack of information in which AI has. My position is that artificial intelligence can reach beyond human limitations in early detection despite having limitations itself.

Therefore, my proposed research question is: Despite limitations, how has artificial intelligence contributed to the early detection of Alzheimer's disease?

Early detection of Alzheimer’s is extremely crucial for effective intervention and disease management. AI plays a pivotal role in neuroimaging, which is where doctors take photos of the brain to detect if a brain shows any symptoms of AD. The article, Artificial intelligence in Alzheimer's diagnosis: A comprehensive review of biomarkers, neuroimaging, and machine learning applications, mainly reviews how AI has improved neuroimaging for magnetic resonance imaging (MRI), positron emission imaging (PET), and computed tomography (CT). MRI utilizes magnetic fields and radio waves to “analyse structural changes in the brain”. PET scans utilize radioactive liquid tracers to assess “metabolic changes and amyloid-beta or tau protein deposition . CT scans utilize X-rays to create images, providing “critical information on brain atrophy.” Traditionally, specialized radiologists will analyze and interpret the scans to look for any abnormalities in the brain. The downside to this includes human error and time consumption. However, AI algorithms such as convolutional neural networks, which analyze visual data, can help speed up this process of data analysis while being accurate. More importantly, these algorithms can “process complex imaging datasets, identifying subtle abnormalities that might be overlooked by traditional methods” (Cretu 2024). This allows for more precision in diagnosis, which is crucial for patient care and treatment. This source agrees with my position, AI is a positive tool for early detection. In my position essay, I can discuss more in detail the specificity of convolutional neural networks and its pivotal role in data analysis. However, the downside to this advancement is the lack of explainability in AI. Typically, AI models will take in data and spit out an answer without really explaining why. Therefore, this can be one of the cases in which limitations have to be addressed in order to ensure radiologists are on board with the process of analysis. Lastly, this source is a peer-reviewed article, which can help me build credibility.

Despite AI revolutionizing detection, one issue is the lack of explainability in artificial intelligence. In the article, Exploration of Explainable AI with Deep Learning Model for Early Detection of Alzheimer's Disease, researchers employ explainable AI techniques that give reasoning to predictions: an image classification AI model, the InceptionV3 model. Why is the explainability of AI models important? Researchers describe artificial models as “black boxes,” making it difficult for clinicians to understand how decisions are made. This lack of transparency limits clinical adoption and trust.” The study thoroughly explains the process of one of the EAI (explainable artificial intelligence) models, LIME (Local Interpretable Model-agnostic Explanations). First, LIME will generate data from MRI scans and track the changes in prediction. Lastly, it will construct a more interpretable model to “determine which features (e.g., certain brain areas) have the most influence on the model's choice. This helps physicians comprehend and accept the AI's predictions by illuminating the image's most important diagnostic features” (Vernekar). This source definitely agrees with my position as it mentions research that improves one of the limitations in AI, explainability. I think this article will greatly benefit my position essay since this study doesn’t just talk about how artificial intelligence can help detection. Instead, the study realizes a flaw in artificial intelligence, and provides the solution to clarify and improve AI. Models like these will definitely be used in the future to battle AI limitations. This article is where I derived my proposal prompt. Therefore, I will be utilizing this source heavily. I can also find other sources similar to this one that talk about the importance of explainability. I believe this will build on the credibility within my argument and further prove my claim.

Surprisingly, another approach to diagnosis is smartphones. The article, Unveiling New Strategies Facilitating the Implementation of Artificial Intelligence in Neuroimaging for the Early Detection of Alzheimer’s Disease, also reviews AI detection advancements in AD and mentions the Skirrow et al. study. This study utilized speech and language patterns along with AI systems for detection of early signs of AD. In the study, 133 patients were “employed daily story recall tasks administered via smartphones”, and then AI systems analyzed the data. As a result, “AI systems demonstrated promising capabilities in predicting MCI and mild AD with a high degree of accuracy” (Etekochay). I think this article agrees with my position as AI is perceived as positive and can open early detection accessibility for patients. This article is also a credible source because it derived from a peer reviewed academic journal, The Journal of Applied Behavioral Science. Furthermore, this article fits with the academic conversation because instead of indulging in traditional neuroimaging through scans, the article describes a different approach in diagnosis with common technology such as smartphones. I can use this in my proposal to further build on the benefits of detection in artificial intelligence through a different perspective. Additionally, there may be some limitations such as limited data from smart phones which I can explore in my proposal.

Another way of diagnosis that research has focused on is ocular biomarkers. In general, biomarkers are measurable substances in the body that can specify a health condition or a biological state. In this case, ocular biomarkers is a measurable characteristic within the eye that can show a presence of Alzheimer's Disease. In the article, Potential Ocular Biomarkers for Early Detection of Alzheimer’s Disease and Their Roles in Artificial Intelligence Studies, study shows that retinal changes and eye fluid biomarkers can help detect AD extremely early because retina and optic nerve are directly related to the central nervous system. Artificial intelligence can process the ocular data and notice subtle changes in the retina before symptoms are too visible. However, ocular biomarker AI analysis still requires further research and testing. The article mentions one of a model’s flaws: “AI research in ophthalmology does not have a large number of datasets that can be used to train AI. This reduces the accuracy and sensitivity of AD detection” (Chaitanuwong). I believe this article is extremely beneficial to my proposal essay because it introduces a different perspective of AI in detection with relations to biomarkers. Biomarkers are a common way in discovering brain atrophy which can be data that AI may have to analyze. In this source, the specific biomarker utilized is ocular biomarkers. However, this is not the only biomarker used to test for AD symptoms. In my proposal essay, I can discuss the potential limited data of specific biomarkers, and how the limitation is addressed. Additionally, the studies mentioned support my position as they introduce another type of data AI can analyze, which can demonstrate the vast broadness of what AI can help with in early detection.

Ultimately, all these sources reveal the commonality that artificial intelligence has a huge positive impact on early detection in Alzheimer's disease. Each source demonstrates a diversity of help in which AI offers: analysis of scan data, biomarker data, explainable AI for detection, and accessibility of detection with smartphones. Each research mentioned is on a completely different part of the AI-in-early-detection-of-AD spectrum. These are only just some of the positive impacts artificial intelligence has brought to early detection. However, limitations are still present in the powerful tool of AI. For instance, the research I have found also mentions some of the biggest limitations of AI. This includes the lack of explainability in decisions (Vernekar) and limited data sets that models can test with (Crețu). It is important for these limitations to be addressed because AI advancement may start to slow down or plateau. Therefore, to further enrich the academic conversation, I can discuss the emerging advancements in AI that have targeted these limitations. By the end of my position essay, I aim to reason that despite limitations in explainability and datasets, AI is still an extremely valuable tool for early detection of Alzheimer’s Disease.

References

• Alzheimer's Association. (n.d.). 2025 Alzheimer's disease facts and figures.
• Chaitanuwong, P., et al. (2023). Potential Ocular Biomarkers for Early Detection of Alzheimer’s Disease.
• Crețu, O. C., et al. (2024). Artificial intelligence in Alzheimer's diagnosis.
• Etekochay, M. O., et al. (2024). AI strategies in neuroimaging.
• Vernekar, S. R., & Kumar, S. (2024). Explainable AI for early Alzheimer’s detection.