Matched-pair design calls for the paired $t$-test

Suppose $n$ randomly chosen persons were enrolled to examine whether different skin creams A and B have different effects on the human body. Cream A was applied to one of the randomly chosen arm of each person, cream B to the other arm. Response is measured as a continuous variable. Which of the following tests is most appropriate to examine the difference?