Advances in information technology have amplified the need for sophisticated tools to analyze complex functional data. Functional Data Analysis meets this demand by emphasizing the selection of key features. This process reduces the data’s dimensionality, making it easier to interpret and draw meaningful conclusions. This paper introduces the common support function, a novel method that leverages quantile and expectile estimation to identify subdomains where two stochastic processes differ the most. By capturing differences in mean and variance, the method highlights regions of minimal overlap between classes, enabling effective domain selection and shape analysis. This tool is applied to real ECG data to identify critical segments associated with myocardial infarction. We were able to pinpoint subintervals that align with medically relevant features such as the QRS complex and ST segments.