Optical coherence tomography and visual evoked potentials: which is more sensitive in multiple sclerosis?