Comparison of speech-evoked v tone-evoked P300 response: implications for predicting outcomes in patients with traumatic brain injury.

摘要

The P300 response is a cognitive event-related potential recorded over the scalp. The tone-evoked P300 response has been used to predict outcomes of patients with brain injury. However, it may lead to false predictions because some normal people have a very small tone-evoked P300 response. It is hypothesized that speech may generate a more robust P300 response than tones. A voice-generator prototype was designed for this study. The rare speech signal was the word "mommy" in a female voice. The common signal was a 1000-Hz tone. Twenty-two normal adults (11 males, 11 females; age range, 18-60 yr) were tested for both speech-evoked and tone-evoked P300 responses. Speech-evoked P300 responses had significantly larger amplitudes (mean, 12.1 microV) than the tone-evoked responses (mean, 5.9 microV; P < 0.0001). Six subjects with brain injury were also tested using the same protocol: two subjects with severe brain injury showed no response to either stimulus. Both died within 1 wk after the testing. Although two subjects with moderate brain injury could not complete the testing because of agitated behavior, two other subjects with mild traumatic brain injury showed a larger speech-evoked than tone-evoked P300 response. The speech-evoked P300 response may be promising in predicting outcomes of patients with brain injury.