Towards an optimal paradigm for simultaneously recording cortical and brainstem auditory evoked potentials