Convergence of Accelerated Brain Volume Decline in Normal Aging and Alzheimer's Disease Pathology.

Background: Age represents the largest risk factor for Alzheimer's disease (AD) but is typically treated as a covariate. Still, there are similarities between brain regions affected in AD and those showing accelerated decline in normal aging, suggesting that the distinction between the two might fall on a spectrum. Objective: Our goal was to identify regions showing accelerated atrophy across the brain and investigate whether these overlapped with regions involved in AD or where related to amyloid. Methods: We used a longitudinal sample of 137 healthy older adults from the Alzheimer's Disease Neuroimaging Initiative (ADNI), who underwent magnetic resonance imaging (MRI). In addition, a total of 79 participants also had longitudinal positron emission tomography (PET) data. We computed linear-mixed effects models for brain regions declining faster than the average to investigate variability in the rate of change. Results: 23 regions displayed a 0.5 standard deviation (SD) above average decline over 2 years. Of these, 52% overlapped with regions showing similar decline in a matched AD sample. Beyond this, the left precuneus, right superior frontal, transverse temporal, and superior temporal sulcus showed accelerated decline. Lastly, atrophy in the precuneus was associated with increased amyloid load. Conclusions: Accelerated decline in normal aging might contribute to the detection of early signs of AD among healthy individuals.

Almanac - Retrieval-Augmented Language Models for Clinical Medicine.

BACKGROUND: Large language models (LLMs) have recently shown impressive zero-shot capabilities, whereby they can use auxiliary data, without the availability of task-specific training examples, to complete a variety of natural language tasks, such as summarization, dialogue generation, and question answering. However, despite many promising applications of LLMs in clinical medicine, adoption of these models has been limited by their tendency to generate incorrect and sometimes even harmful statements. METHODS: We tasked a panel of eight board-certified clinicians and two health care practitioners with evaluating Almanac, an LLM framework augmented with retrieval capabilities from curated medical resources for medical guideline and treatment recommendations. The panel compared responses from Almanac and standard LLMs (ChatGPT-4, Bing, and Bard) versus a novel data set of 314 clinical questions spanning nine medical specialties. RESULTS: Almanac showed a significant improvement in performance compared with the standard LLMs across axes of factuality, completeness, user preference, and adversarial safety. CONCLUSIONS: Our results show the potential for LLMs with access to domain-specific corpora to be effective in clinical decision-making. The findings also underscore the importance of carefully testing LLMs before deployment to mitigate their shortcomings. (Funded by the National Institutes of Health, National Heart, Lung, and Blood Institute.).