Cortical evoked potentials of speech in noise for normal hearing listeners

Normal hearing listeners have a wide range of speech recognition outcomes in noise, and it is not clear why this variation occurs. In order to investigate the cause of this variation between listeners we need to use objective tests that can predict speech in noise performance. This study evaluates cortical evoked potentials for this purpose, more specifically, the acoustic change complex (ACC), which is an event related potential (ERP) that reflects cortical auditory processing. The aim is to see if speech perception performance in noise is reflected in the ACC. Thirty normal hearing listeners (aged 19-38 years) were tested on their speech in noise abilities and their ACC responses to random continuous sequences of vowels and fricatives at three different noise levels. The ACC stimuli consisted of four vowels, four fricatives and silence, each lasting 300-400 ms, in quiet, at -3 dB SNR and +4 dB SNR. Participants listened to 20 minutes of these sequences at each noise level whilst their ACC was measured using a Biosemi EEG system. Their speech in noise abilities were assessed using a sentence recognition task at -3, -6 and -9 dB SNR. Preliminary results show that with increasing noise level the amplitude of the ACC decreases and the latency increases. Furthermore, the latencies of the P1 and N1 peaks decrease with increasing speech in noise ability for certain stimulus pairs at +4 dB SNR. It therefore appears that the ACC response has the potential to be used to predict speech in noise performance in normal hearing listeners.