Human Time Perception - Predictable visual stimuli are perceived earlier than unpredictable events

Abstract

What is it that enables us to timely react to visual events despite the significant processing delays within the visual system? This delay is estimated to be already about 100ms in higher visual areas, a delay that is relevant if one needs to initiate fast reactions, such as catching a ball in flight or initiating an escape. To compensate for delays in motor behavior, the brain employs predictive mechanisms. We aimed to investigate whether predictability of a visual stimulus not only affects behavior but also the time of perceived stimulus onset in humans. Specifically, we hypothesized that predictable visual stimuli have an earlier perceived onset than nonpredictable stimuli do. Our approach was the following: Subjects viewed streams of individual letters with a 1000ms standard interval between letters. This sequence of letters was either in alphabetical order, and thus predictable, or alternatively, the last letter of a sequence was chosen at random and thus not predictable. In each trial, subjects had to indicate whether or not the last letter agreed with the alphabetical order. Moreover, subjects had to estimate whether the duration of the last test interval, which was of varying length, was either longer or shorter than the standard interval. Varying the length of the last interval allowed us to estimate the point of subjective equivalence (PSE) between test and standard intervals. As we expected predictable letters to be perceived earlier, the PSE should be larger in predictable sequences than in nonpredictable ones. In other words, the test interval would need to be relatively longer in order to compensate for the earlier perceived onset of predictable as compared to unpredictable letters. Measurements of the PSEs confirmed our expectations and suggest that predictable visual stimuli are perceived earlier than non-predictable ones. Hence, even the perceptual system is compensating for delays in sensory information processing, allowing us to establish a timely perception of our environment. To shed light on the related neuronal correlates of delay compensation, we performed a magnetoencephalography (MEG) study while subjects performed the same task. We intended to investigate relative temporal differences between predictive and non-predictive visual evoked potentials (VEPs) in our MEG study. Analysis of MEG data suggests that participants did generate predictions. This is indicated by a late signal difference between the non-predictable and the predictable stimuli. However, we found no evidence for predictable stimuli evoking an earlier (or higher) peak of VEPs than unpredictable stimuli. Ultimately, it remains open whether the earlier perception of predictive vs. non-predictive stimuli is mediated through sensory prediction, through an interplay of prediction and postdictive perceptual evaluations, or through prediction and some yet unknown delay compensation mechanism.