Comparative Analysis of State-of-the-Art Foundation Models for Sleep Analysis Under Channel Reduction

This paper evaluates six sleep staging foundation models under varying levels of PSG channel reduction using the Multi-Ethnic Study of Atherosclerosis dataset. The research addresses the high cost and burden of full PSG montage, investigating how performance degrades when reducing sensor channels for at-home or long-term monitoring applications. The study provides benchmark comparisons across different model architectures under constrained channel conditions.