Page 45 - It' about time: Studying the Encoding of Duration

Chapter 2 In the current study, we did not find any evidence for a decrease in the duration after-effect with increased adapter-test distance. While this indicates a strong role for later visuals areas, it does not fully exclude a role for earlier areas. Many adaptation after-effects for more complex stimulus properties (i.e. motion, numerosity, faces) with processing loci in higher-level areas have been shown to also include a low-level adaptation component (Afraz & Cavanagh, 2008; Kovács et al., 2005, 2007). One classic example is the contribution of adaptation in V1 in motion after-effects adaptation (Kohn & Movshon, 2003). The fact that we do not find any evidence for such a lower-level contribution in the duration after-effect might reflect a lack of sensitivity in the method deployed in this experiment. Incorporating neuroscientific methods such as fMRI might provide a more appropriate way to address this question. Despite this nuance about a lower-level contribution, it is safe to conclude that the main locus for duration encoding lies outside earlier visual areas. The results reported here are in strong contrast with the result of studies investigating the effect of adaptation to other temporal and non-temporal stimulus features on perceived duration (Johnston et al., 2006; Ortega et al., 2012; Zhou et al., 2014). In general, these studies report strong spatial selectivity indicating an origin in early visual areas (e.g. LGN, V1). This suggests that the after-effects reported by these studies and the duration after-effect result from distinct mechanisms located at different stages of visual processing. This, however, raises the question of why distinct mechanisms exist. One answer to this question is to assume a hierarchical structure to the processing of duration (Heron et al., 2013; van Wassenhove, 2009). Models proposing such a structure focus on the idea that temporal information is intrinsically present in any sensory signal (Buonomano, 2000; Buonomano & Maass, 2009; Karmarkar & Buonomano, 2007; van Wassenhove, 2009). These patterns of different states of the network can then be extracted to form a more explicit duration code (e.g. via channel-based encoding) so that it can be integrated across the senses. The resulting multimodal representation can then be stored and interact with other information to inform subsequent behavior. Assuming that the different after-effects reflect different hierarchical steps in duration processing could explain the observed differences in spatial selectivity of these after-effects. We would suggest that adaptation to temporal frequency and non-temporal factors occurs earlier in duration processing by causing modulations in the states of the network. The duration-selective mechanisms then act upon these 44