Page 118 - It' about time: Studying the Encoding of Duration

responses eventually returning to baseline (Heron et al., 2012; Chapter 1, Figure 4). This pattern of results is consistent with temporally localized sensory adaptation, which only affects the subsequent encoding of similar durations. In contrast, a decision level bias as a result of the recalibration of responses would occur at all adaptation durations regardless of the difference between the adapted and tested duration. As such, these results argue against a purely decisional bias account of the DAE. A final argument against the presence of a decision level bias in the DAE comes from the reported dissociations between the perceived duration of a stimulus, and the DAE measured following adaptation to that same stimulus (Heron et al., 2013, Chapter 4). If decision level biases reflect recalibration of responses based on earlier evaluations of sensory input (Mather & Sharman, 2015), one would expect a decision level bias to occur based on the durations available for decision (i.e. the durations perceived by the observer). In Chapter 4, we adapted observers to an illusion-inducing stimulus and found no evidence that they adapted to the duration corresponding to the perceived duration of that stimulus. Instead, we found that the DAE reflected adaptation to the temporal distance between the on- and offset of an event (onset-offset duration). Similarly, Heron and colleagues (2013) adapted observers to multi- modal stimuli for which the perceived duration of the visual component was shifted towards the duration of the auditory component. In contrast to this shift in perceived duration, adaptation to these same stimuli resulted in adaption to the duration of the visual component and not the duration perceived by the observer. Both these studies demonstrate a dissociation between perceived duration and the duration adaptation reflected by the DAE. These patterns of results are unlikely to result from decision level bias as the information available for decision (and recalibration) and the duration to which observers adapt are 6 different. While the arguments presented above provide converging evidence against a strong role for decision level bias in the DAE, they do not exclude the occurrence of a decision level biases from contributing to the DAE per se. One approach to further address this issue would be to design experimental methods that aim to reduce decision level bias (Mather & Sharman, 2015; Storrs, 2015). For example, by using altered variations on the nAFC paradigm to manipulate decision level bias (García-Pérez & Alcalá-Quintana, 2013; Jogan & Stocker, 2014; Morgan, 2013, 2014). Alternatively, paradigms could aim to General discussion 117