Grant Number:	2R01CA236793-06A1 Interpret this number
Primary Investigator:	Whitney, David
Organization:	University Of California Berkeley
Project Title:	Isolating and Mitigating Sequentially Dependent Perceptual Errors in Clinical Visual Search
Fiscal Year:	2025

Remote-store-and-forward teledermatology has grown exponentially in popularity as an efficient, accurate, and cost-effective way to improve the health and well-being of millions of patients around the world. In remote store- and-forward teledermatology, clinicians are asked to recognize and classify images of skin lesions and make visual judgments about lesion features and malignancy. Dermatologists performing screening in this paradigm can examine hundreds of images, seeing them one after the other. A central underlying assumption of this task is that dermatologists’ percepts and decisions about a current image are completely independent of prior events. Recent results show that this is not true: our perception and decisions are strongly biased by our past visual experience. Although serial dependencies were proposed to be a purposeful mechanism to achieve perceptual stability in natural vision, serial dependencies in remote store-and-forward teledermatology could play a crucial and deleterious role in cancer screening. For example, a malignant lesion could be classified as benign depending on the content of the previously seen image. Given the importance and impact of serial dependencies in dermatological settings, we plan to (1) establish, (2) identify, and (3) mitigate the conditions under which serial effects determine percepts and decisions in a remote store-and-forward teledermatology setting. In Aim 1, based on preliminary and pilot data that indicate a clear detrimental role of serial dependencies in clinical visual judgments, we plan to test the full impact of serial dependencies on lesion recognition, including four common tasks: malignancy classification, lesion symmetry perception, lesion homogeneity perception, and lesion border discrimination. In Aim 2, we plan to identify the specific boundary conditions under which visual serial dependence impacts visual recognition in store-and-forward teledermatology. In Aim 3, we will use the boundary conditions identified in Aim 2 to propose a series of task and stimulus manipulations to mitigate the deleterious effects of visual serial dependence on lesion recognition. As a result of these manipulations, performance should improve in measurable ways (including sensitivity, specificity, accuracy, and d’). Aim 3 is particularly crucial because it will allow us to propose new guidelines that will greatly improve lesion recognition in remote store- and-forward teledermatology screening. Taken together, the proposed studies in Aim 1, 2, and 3 will allow us to establish, identify, and mitigate the deleterious effect of serial dependencies in remote store-and-forward teledermatology tasks, which could have a significant impact on the health and well-being of skin cancer patients around the world.