The Role of the Human Auditory Middle Latency Response in Auditory Novelty Detection

Resumen

One way of structuring the huge amount of sound input reaching the human ear, is extracting sound objects, which are formed by stimuli having a certain regularity in common. Mismatching stimuli are detected pre-attentively and can trigger an involuntary attention switch towards them. The auditory-event related potential which reflects the process of detecting mismatching sounds is called mismatch negativity (MMN), usually peaks at 150 to 250 ms from stimulus onset and has bilateral sources in auditory and prefrontal cortex. MMN is elicited by deviants violating auditory regularities, like in a simple “oddball paradigm”, which is composed of unchanging repetitive “standard” sounds and rare, randomly occurring, deviating sounds (so called deviants), as well as by violations of more complex auditory regularities. Recently it has been shown that also the middle latency response (MLR), which is an earlier auditory evoked potential (AEP), reflects the occurrence of deviating sounds in an oddball paradigm. The objective of the present thesis was to examine the role of the middle latency response in the auditory deviance detection system. The first study showed that lower intensity deviants of an oddball paradigm elicited a slight negativity at the transition from the Na to the Pa wave, in comparison to the response elicited by physically identical standard stimuli. In addition, an MMN was elicited. In the second study a hypercomplex invariance, more concretely, a feature-conjunction paradigm with two types of standard stimuli, each with a distinct combination of stimulus frequency and stimulus source location, and two types of deviant stimuli, each with the frequency of one standard stimulus, and the location of the other, were presented. In order to compare the results with MLRs elicited by stimuli of a simple auditory regularity, an additional simple oddball paradigm with frequency deviants was presented. The Nb wave of the MLR was enhanced in response to frequency deviants compared to standard stimuli of the simple oddball paradigm. However, comparison of the MLRs to deviants and standards of the feature-conjunction paradigm yielded no differences. An MMN was elicited in both paradigms. In the third study the application of a variation of the multi-feature paradigm for MLR studies and the MLR in response to frequency-intensity double deviants were probed. Frequency-intensity double deviants elicited a significant enhancement of the MLR, which was as large as the sum of the enhancements elicited by the frequency and intensity single deviants. The results of the present thesis suggest that the early deviance detection at the level of the MLR occurs only for simple auditory regularities, as in the case of feature repetitions in the simple oddball paradigm, or in the multi-feature paradigm, where formation of the standard trace does not require extracting feature-combinations. Furthermore, the results suggest that deviations in frequency and intensity are processed independently from each other. Based on evidence from the present thesis and from other studies, we conclude that the regularity encoding and deviance detection of stimuli presented in more complex auditory regularities than the simple oddball or the multi-feature paradigm require higher-order brain mechanisms than those reflected in the MLR. This goes in line with the hypothesis of a hierarchically organized auditory novelty system. Concerning the cellular mechanisms underlying auditory deviance detection, it has been proposed that stimulus-specific adaptation (SSA) to stimulus probabilities observed in animal auditory subcortical and cortical structures could be the single neuron correlate of the deviance-related activity in the human AEP due to its similar characteristics to MMN. However, since the latencies of the MLR better match the latencies of SSA, it is probable that the deviance-related modulations in the MLR represent a more direct correlate of the early cellular SSA, than MMN.