Page 119 - It' about time: Studying the Encoding of Duration
P. 119

                                Chapter 6  measure adaptation-associated changes in sensitivity instead of changes in the PSE, which should not be susceptible to decision level changes (Clifford, Wyatt, Arnold, Smith, & Wenderoth, 2001; Kohn, 2007; Morgan, Chubb, & Solomon, 2006). While these methods provide valuable tools to manipulate, measure, and reduce the contribution of decision level bias, it is not always possible or practical to transfer these methods to the domain of temporal judgements. For example, some improved variations of nAFC tasks require concurrent presentation of multiple reference-test pairs which is not possible when studying duration (Jogan & Stocker, 2014; Morgan, 2014). Instead of using an nAFC, a more valuable approach could be to use experimental methods that allow for the dissociation and quantification of changes in perceptual evidence (perceptual bias) and changes in decision criterion (decision level bias). One example of such a method would be to quantify different changes in perceptual decision-making by modeling decision-making data using a drift diffusion model (DDM). DDMs assume that a dichotomous decision-making process is characterized by the accumulation of evidence until a certain decision threshold is reached (Forstmann, Ratcliff, & Wagenmakers, 2016; Ratcliff, 2014; P. L. Smith, 2000). The higher the quality of the sensory evidence, the higher the rate at which evidence can be accumulated (drift rate). In addition, changes in decision threshold can also affect the probability that a certain decision is reached. This distinction between drift rate and decision threshold is analogous to the perceptual and decision level biases described earlier. If duration adaptation causes test stimuli to be perceived as more distinct from the adapted duration (perceptual bias) it should lead to a predictable increase in discriminability for certain perceptual decisions (i.e. increase in drift rate). Conversely, decision level biases would manifest themselves as shifts in the relative distance to the two decision thresholds. By modeling both decisions and the associated response times, drift diffusion models can be used to understand the relative contribution of changes in drift rate and changes in thresholds to observed changes in decision making. By taking advantage of this characteristic, drift diffusion models could provide new insight into the prevalence of both perceptual and decision biases in the DAE. Relating behavioral outcomes to neurophysiological processes Throughout this dissertation, I discussed two main categories of evidence that support the notion that duration-tuned responses play a role in our capacity 118 


































































































   117   118   119   120   121