Support for the interruption and protection hypotheses of prepulse inhibition of startle: evidence from a modified Attention Network Test.