Stimuli generated by a person’s own willed actions generally elicit a suppressed neurophysiological response than physically identical stimuli that have been externally generated. This phenomenon, known as sensory attenuation, has primarily been studied by comparing the N1, Tb and P2 components of the event-related potentials (ERPs) evoked by self-initiated vs. externally generated sounds. Sensory attenuation has been implicated in some psychotic disorders such as schizophrenia, where symptoms such as auditory hallucinations and delusions of control have been conceptualised as reflecting a difficulty in distinguishing between internally and externally generated stimuli. This thesis employed a novel paradigm across five experiments to investigate the role of agency and contingency in sensory attenuation. The role of agency was investigated in in Chapter 2. In Experiment 1, participants watched a moving, marked tickertape while EEG was recorded. In the active condition, participants chose whether to press a button by a certain mark on the tickertape. If a button-press had not occurred by the mark, then a tone would be played one second later. If the button was pressed prior to the mark, the tone was not played. In the passive condition, participants passively watched the animation, and were informed about whether a tone would be played on each trial. The design for Experiment 2 was identical, except that the contingencies were reversed (i.e., pressing the button prior to the mark led to a tone). The results were consistent across the two experiments: while there were no differences in N1 amplitude between the active and passive conditions, the amplitude of the Tb component was suppressed in the active condition. The amplitude of the P2 component was enhanced in the active condition in both Experiments 1 and 2. These results suggest that agency and motor actions per se have differential effects on sensory attenuation to sounds and are indexed with different ERP components. In Chapter 3, we investigated the role of contingency in sensory attenuation while using a similar ticker-tape design in Chapter 2. In the Full Contingency (FC) condition, participants again chose whether to press a button by a certain mark on the tickertape. If a button-press had not occurred by the mark, a sound would be played (one second later) 100% of the time (Experiment 3). If the button was pressed prior to the mark, the sound was not played. In the Half Contingency (HC) condition, participants observed the same tickertape; however, if participants did not press the button by the mark, a sound would occur 50% of the time (HC-Inaction) while if the participant did press the button, a sound would also play 50% of the time (HC-Action). In Experiment 4, the design was identical, except that a button-press triggered the sound in the FC condition. The results were consistent across both Experiments in Chapter 3: while there were no differences in N1 amplitude across the FC and HC conditions, the amplitude of the Tb component was smaller in the FC condition when compared to the HC-Inaction condition. The amplitude of the P2 component was also smaller in the FC condition compared to both the HC-Action and HC-Inaction conditions. The results suggest that the effect of contingency on neurophysiological indices of sensory attenuation may be indexed by the Tb and P2 components, as opposed to the more heavily studied N1 component. Chapter 4 also investigated contingency but instead used a more ‘traditional’ self-stimulation paradigm, in which sounds immediately followed the button-press. In Chapter 4, participants observed a fixation cross while pressing a button to generate a sound. The probability of the sound occurring after the button-press was either 100% (active 100) or 50% (active 50). In the two passive conditions (passive 100 and passive 50), sounds generated in the corresponding active conditions were recorded and played back to participants while they passively listened. In contrast with the results of Chapter 3, the results of Chapter 4 showed both the classical N1 suppression effect, and also an effect of contingency of the N1, where sounds with a 50% probability generated higher N1 amplitudes compared to sounds with 100% probability. In contrast, Tb amplitude was modulated by contingency but did not show any differences between the active and passive conditions. The results of this study suggest that both sense of agency and sensory contingency can influence sensory attenuation, and thus should be considered in future studies investigating this theoretically and clinically important phenomenon.