The role of expectations on affective sound processing: behavioural and neural correlates

Abstract

Theoretical frameworks and empirical evidence in the last two decades have shown that prior expectations about the upcoming stimulus can shape the perception when the stimulus arrives. How expectations influence emotional responses to the stimulus is, however, relatively less understood. In this thesis, using behavioural and neural measures of electroencephalography (EEG) and intracranially recorded local field potentials (LFPs) from human subjects, I explore the role of expectations on the processing of affective sounds. In Chapter 2, the neural basis of expectation is first established using EEG. Two visual cues were used to elicit the expectation of either neutral or aversive sounds. An event related potential just before the onset of upcoming stimuli, called Stimulus Preceding Negativity (SPN), is measured to index the expectation of an upcoming stimulus. Although a robust measure of SPN could be observed for the expectation of both aversive and neutral sounds, no difference between the two was observed indicating no relation between the magnitude of SPN and valence of sounds. Source localization of SPN, using multiple sparse priors algorithm revealed a network of brain areas including the anterior insula, inferior frontal gyrus, temporal cortex, supplementary motor area (SMA) and thalamus. A limitation of the first experiment was that no behavioural measure of expectation of valence was recorded. It is likely that there is variation across subjects in the expectation and perceived valence after the stimulus onset. The second experiment (Chapter 3), also a cued paradigm as above, addressed this limitation by using subjective measures of expectations before the sound onset and aversive ratings after sound offset as reported by the subjects. Mediation analysis between perceived ratings following sound onset and expectation ratings confirmed a mediator role of expectation/predictions in the aversive experience – an expectation for aversive sounds translated into a more aversive experience, and an expectation for less aversive sounds translated into a less aversive experience once the sound was heard. Exploratory analyses showed that subjects whose perceived aversiveness shifted in the direction of expectation displayed a stronger SPN. Moreover, this effect was seen for aversive

but not for neutral sounds. Additionally, activity in the alpha-beta band during encoding of the predictive cues was associated with the precision of subjective expectancy. In summary, the data highlight the importance of measuring behavioural/subjective correlates of expectation and perceived aversiveness. This may be particularly important when the cues-contingencies are not explicitly disclosed and when using emotional (subjective) stimuli, as there is bound to be high inter-individual variability both in learning rates and stimulus appraisal. Expectations about the upcoming stimulus can be formed based on different sources. For example, it could be based on information from other people (that is, social source) or expectations can be formed based on personal experiences with the world (conditioned source). In the third and last experiment (Chapter 4), the behavioural, physiological and neural basis of social and conditioned expectations are measured. Using a cued paradigm, subjects formed expectations of the upcoming stimulus either based on social information or their own conditioning experiences. As in the experiment in Chapter 3, subjects rated their expectation prior to the stimulus onset and their perceived aversiveness following the onset of the stimulus. The data again show that the perceived aversiveness shifted in the direction of expectations for both the social and conditioned cues. Further, physiological and autonomic responses also shifted in the direction of expectations. Recordings from LFPs in a group of epileptic patients undergoing neurosurgical evaluations for the locations of their seizure foci show expectation-based changes during sound perception in a widespread network including temporal cortex, anterior cingulate and orbitofrontal cortices, inferior frontal gyrus, and insula. Collectively, the research presented in this doctoral thesis show expectations can and do alter the processing of aversive sounds at the behavioural, somatic, and neural levels.