Yiwen Li

This project represents the first systematic assessment of the US Food and Drug Administration’s postmarket surveillance of legally marketed artificial intelligence and machine learning based medical devices. We focus on the Manufacturer and User Facility Device Experience database—the FDA’s central tool for tracking the safety of marketed AI/ML devices. In particular, we evaluate the data pertaining to adverse events associated with approximately 950 medical devices incorporating AI/ML functions for devices approved between 2010 through 2023, and we find that the existing system is insufficient for properly assessing the safety and effectiveness of AI/ML devices. Inparticular,we make three contributions: (1) characterize the adverse event reports for such devices, (2) examine the ways in which the existing FDA adverse reporting system for medical devices falls short, and (3) suggest changes FDA might consider in its approach to adverse event reporting for devices incorporating AI/ML functions.

When two repeating stimuli activate the same neuronal population, a decreased overall neural response is observed. This neurophysiological response is detectable by functional magnetic resonance imaging, and has been termed fMRI adaptation. Following this logic, when two tactile events are repeated on exactly the same region of skin, all neurons that have a strictly somatotopic response should reduce their activity. Here we used fMRI adaptation to address the issue of reference frames for touch. In particular, we asked if there exist neuronal populations that adapt to stimulation that repeats over distinct regions of skin when some other aspect of spatial coding of touch is identical (e.g., homologous finger stimulated between the two hands). We examined this question for SI and SII, measuring the blood-oxygen level-dependent (BOLD) effect to touches delivered in sequence within or between hands, to homologous or non-homologous fingers. The results showed stronger adaptation (i.e., a lower BOLD effect) at the level of SI when stimulation repeated over same (i.e., homologous) in comparison to different (i.e., non-homologous) fingers, regardless of whether stimulation occurred within or between hands. This pattern of results is less clear in SII. These results suggest a bilateral representation for touch in human SI, in agreement with neurophysiological data on monkeys.