A three-tone sinusoidal replica of a naturally produced utterance was identified by listeners, despite the readily apparent unnatural speech quality of the signal. The time-varying properties of these highly artificial acoustic signals are apparently sufficient to support perception of the linguistic message in the absence of traditional acoustic cues for phonetic segments.

Children with learning problems often cannot discriminate rapid acoustic changes that occur in speech. In this study of normal children and children with learning problems, impaired behavioral discrimination of a rapid speech change (/dalpha/versus/galpha/) was correlated with diminished magnitude of an electrophysiologic measure that is not dependent on attention or a voluntary response. The ability of children with learning problems to discriminate another rapid speech change (/balpha/versus/walpha/) also was reflected in the neurophysiology. These results indicate that some children's discrimination deficits originate in the auditory pathway before conscious perception and have implications for differential diagnosis and targeted therapeutic strategies for children with learning disabilities and attention disorders.

Behavioral perceptual abilities and neurophysiologic changes observed after listening training can generalize to other stimuli not used in the training paradigm, thereby demonstrating behavioral “transfer of learning” and plasticity in underlying physiologic processes. Nine normal-hearing monolingual English-speaking adults were trained to identify a prevoiced labial stop sound (one that is not used phonemically in the English language). After training, the subjects were asked to discriminate and identify a prevoiced alveolar stop. Mismatch negativity cortical evoked responses (MMN) were recorded to both labial and alveolar stimuli before and after training. Behavioral performance and MMNs also were evaluated in an age-matched control group that did not receive training. Listening training improved the experimental group’s ability to discriminate and identify an unfamiliar VOT contrast. That enhanced ability transferred from one place of articulation (labial) to another (alveolar). The behavioral training effects were reflected in the MMN, which showed an increase in duration and area when elicited by the training stimuli as well as a decrease in onset latency when elicited by the transfer stimuli. Interestingly, changes in the MMN were largest over the left hemisphere. The results demonstrate that training can generalize to listening situations beyond those used in training sessions, and that the preattentive central neurophysiology underlying perceptual learning are altered through auditory training.

A passively elicited cortical potential that reflects the brain's discrimination of small acoustic contrasts was measured in response to two slightly different speech stimuli in adult human subjects. Behavioral training in the discrimination of those speech stimuli resulted in a significant change in the duration and magnitude of the cortical potential. The results demonstrate that listening training can change the neurophysiologic responses of the central auditory system to just-perceptible differences in speech.

The mismatch negativity (MMN) is an automatic cortical evoked potential that signifies the brain's detection of acoustic change. In other words, the MMN reflects the neurophysiologic processes that underlie auditory discrimination. As such, the MMN provides an objective tool for evaluating central auditory mechanisms involved in speech perception.We are using the MMN to study the central auditory processes that encode acoustic changes important for speech perception in 1) normal-hearing adults and children, 2) individuals with impaired auditory systems (including persons with learning disabilities, attention deficit disorders, cochlear implants), and 3) an animal model. Specifically, we have demonstrated that the MMN provides information about the central processing of fine acoustic differences, the neuroanatomic pathways that encode acoustic change, central auditory processing in the presence of peripheral hearing deficits, and central auditory system plasticity, In addition, we have considered methodological challenges associated with measuring the MMN in individual subjects.Several methodological issues-including appropriate stimuli, stimulus presentation variables, the recording protocol and environment, and validation of the MMN in individuals-are discussed.

No abstract

Stevens and Blumstein [Perspectives on the Study of Speech (Erlbaum, Hillsdale, NJ, 1981)] have proposed that the global shape of the CV syllable onset spectrum provides the listener with a primary and contextually invariant cue for place of stop consonant articulation. Contextually variable formant transitions are, in contrast, claimed to constitute secondary cues to place of articulation that, during development, are learned through their co-occurrence with the primary spectral ones. In the two experiments reported here, these claims about the relative importance of the onset spectrum and formant transition information were assessed by obtaining adults' and young children's identifications of synthetic stimuli in which these two potential cues specified different places of articulation. In general, the responses of both adults and children appeared to be determined by the formant transitions of the stimuli. These results provide little support for the claim that sensitivity to the global properties of the onset spectrum (as described by Stevens and Blumstein) underlie place of articulation perception or for Stevens and Blumstein's primary versus secondary cue distinction. Rather, these findings are consistent with the view that dynamic, time-varying information is important in the perception of place of articulation.

The mismatch negativity (MMN) is a passively elicited event-related potential that is extremely sensitive to acoustic stimulus properties. The MMN was characterized in normal adults and school-age children in response to speech stimuli differing minimally in the onset frequency of the second and third formant transitions. The speechevoked MMN consists of a negative waveform at about 230 msec that occurs in response to the deviant stimulus when it is presented in an oddball paradigm. It is absent in response to that same stimulus when presented alone. The MMN was clearly present in all adults and children tested. Using the procedures developed in this study, this event-related potential was found to be robust enough in individual subjects to be considered a potential clinical measure for assessing central auditory function in schoolage children and adults. (Ear Hear 13 3:158-164)