In the below link under (LTI Systems with Random Inputs:) the author writes: 
https://www.probabilitycourse.com/chapter10/10_2_2_LTI_systems_with_random_inputs.php
How do you convolve two functions with different arguments? What specifically does $h(\tau) * R_X(-\tau)$ mean?
I know that $(h * R_X)(\tau) = h(\tau) * R_X(\tau) = \int_{-\infty}^{+\infty}h(\alpha)R_X(\tau - \alpha) d\alpha$, but how does $h(\tau) * R_X(-\tau)$ become the integral in the picture?