Abstract
Reaction times (RTs) are used extensively to study decision-making, including to estimate the speed of decision-making processes. RT, however, is composed of various components (i.e. motor response speed). As only a portion of measured RTs represents decision time, RTs need to be decomposed into their components to estimate decision time; a task that is usually done using drift-diffusion models (DDM; Ratcliff & McKoon, 2008). Here we consider a different approach that relies on duration threshold measurements, defined as the shortest presentation time sufficient to accurately perceive a time-limited stimulus. For example, when a stimulus is noisy, a longer presentation duration will be needed to accumulate enough information for an accurate response. Here, we aimed to test whether duration thresholds can provide information that corresponds to decision times extracted from RTs using DDM.
We measured motion direction discriminations for moving Gabor patches at two contrasts (low: 2.1%; high: 92%) and three sizes (1.2, 2.7- and 6-degrees radius). At low contrast, participants (n = 8) performed better (i.e. lower duration threshold) as size increases, while at high contrast, participants performed worse as size increases. These results replicated previous findings of contrast-dependent spatial suppression (Tadin et al., 2003). The overall RTs with the same stimuli did not reveal the same pattern of results. However, decision times extracted from RTs using a DDM did closely follow duration threshold results. Across individuals, we found that decision times are highly correlated with duration thresholds (r > 0.90).
These results suggest that duration thresholds can be used as a measure of decision time, with an added benefit of not including non-decision processes (e.g., motor speed) or requiring the use of DDMs to extract the decision time component. We are currently extending these results in the context of a static orientation discrimination task while varying stimulus contrast.