Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study evaluates the spatial and temporal fidelity of deep learning models estimating left-ventricular ejection fraction (EF) from echocardiography, specifically analyzing a VideoMAE transformer and a Kinetics-pretrained R(2+1)D CNN. While both models demonstrate strong anatomical relevance, with intersection-over-relevance scores significantly above chance, they fail to accurately localize the clinically critical end-systolic and end-diastolic frames, indicating a temporal blindness. The findings highlight a critical gap in current attribution methods, suggesting that spatially faithful explanations may not guarantee temporal accuracy in video diagnostic models.
Spatially faithful models can still overlook crucial temporal information, raising concerns about the reliability of current attribution methods in medical diagnostics.
Background and Objective: Deep video models estimate left-ventricular ejection fraction (EF) from echocardiography with near-expert accuracy, and post-hoc attribution (Chefer relevance for transformers, Grad-CAM for CNNs) is increasingly used to certify that models"look at the right place."Yet whether these explanations are faithful both spatially and temporally is unaudited. Because EF is defined by the end-systolic (ES) and end-diastolic (ED) frames, a faithful explanation must localize the left ventricle (space) and the decisive frames (time). Methods: We fine-tune two distinct EF regressors on EchoNet-Dynamic -- a self-supervised VideoMAE transformer and a Kinetics-pretrained R(2+1)D CNN -- and audit each with architecture-matched attribution along three axes: intersection-over-relevance (IoR) against LV masks, deletion AUC, and a temporal localization index on ES/ED frames, each relative to chance with per-case 95% CIs over 50 studies. A tubelet-occlusion probe separates attribution failure from model behavior. Results: Both models are anatomically faithful -- IoR 2.91x (VideoMAE) and 1.98x (R(2+1)D) above chance -- yet temporally blind: temporal localization is indistinguishable from chance (0.97--1.00) and no better than random attribution. Occlusion shows the models do not preferentially rely on ES/ED (0.90x chance), so temporal blindness reflects model behavior, not an attribution artifact. Conclusions: Spatial faithfulness does not imply temporal faithfulness. Attribution can certify anatomical grounding while masking that a model ignores the clinically decisive frames -- a caution for XAI-based validation of video diagnostic models and a call for temporally-aware training and evaluation.