This study investigates adaptation of high-frequency cortical responses [>60 Hz; high-gamma (HG)] to simple and complex sounds in human nonprimary auditory cortex. We used intracranial electrocorticographic recordings to measure event-related changes in HG power as a function of stimulus probability. Tone and speech stimuli were presented in a series of traditional oddball and control paradigms. We hypothesized that HG power attenuates with stimulus repetition over multiple concurrent time scales in auditory association cortex. Time-frequency analyses were performed to identify auditory-responsive sites. Single-trial analyses and quantitative modeling were then used to measure trial-to-trial changes in HG power for high (frequent), low (infrequent), and equal (control) stimulus probabilities. Results show strong reduction of HG responses to frequently repeated tones and speech, with no differences in responses to infrequent and equal-probability stimuli. Adaptation of the HG frequent response, and not stimulus-acoustic differences or deviance-detection enhancement effects, accounted for the differential responses observed for frequent and infrequent sounds. Adaptation of HG responses showed a rapid onset (less than two trials) with slower adaptation between consecutive, repeated trials (2-10 s) and across trials in a stimulus block (∼7 min). The auditory-evoked N100 response also showed repetition-related adaptation, consistent with previous human scalp and animal single-unit recordings. These findings indicate that HG responses are highly sensitive to the regularities of simple and complex auditory events and show adaptation on multiple concurrent time scales in human auditory association cortex.

Zsiga and Nitisaroj [(2007)] conducted tone perception experiments to test the Moren and Zsiga [(2006)] hypothesis that the principal perceptual cues to the five-way tonal contrast in Thai are high (H) and low (L) pitch targets aligned to (subsyllabic) moras. Their experiments involved perception of synthetic stimuli: manipulated peaks or troughs (corresponding to pitch targets) connected by line-trajectories. In the present study, the stimuli created are intended to be more naturalistic. The hypothesis is that not only peaks and troughs but also the trajectories between them are informative to and important for perception. Manipulations involve realignment of the peaks and troughs, with concomitant compression or expansion of the original trajectory between them. Native Thai speakers categorize the manipulated tone and give a goodness rating. Preliminary data indicate that some categorizations (e.g., manipulated rising perceived as high) can be explained by information present in the trajectory, and not in the peak/trough. However, theories that advocate movements (i.e., trajectories) as perceptual cues [e.g., Xu (2004)] cannot account for some goodness ratings observed in these experiments. Hence, it appears that a combination of the peak/trough information, as well as the trajectory, is employed in tone perception. Work supported by Dr. Luigi Burzio and Dr. Colin Wilson.

Tone languages use pitch contrastively. For example, Thai has 5 contrastive tones, namely Mid, High, Low, Falling and Rising. The latter two ('contour' tones) show greater ranges of f0 values, while the first three ('level tones) show smaller ranges of f0 values.