Divided attention disrupts perceptual encoding during speech recognition.