The Student-t distribution was chosen over the Gaussian distribution because its heavier tails better capture outliers and extreme physiological events in clinical data.