Purchase this article with an account.
Mingbo Cai, David Eagleman; Dissecting the neural network of duration perception with fMRI. Journal of Vision 2014;14(10):1140. doi: 10.1167/14.10.1140.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Traditional neural models of time perception often assume a dedicated amodal brain system for encoding time. In contrast, some recent models suggest that time may be intrinsically encoded in sensory cortices that also code for other aspects of stimuli. The role of the sensory cortices in time perception is often overlooked in functional magnetic resonance imaging (fMRI) studies. To date, few studies have attempted to decode temporal features of sensory stimuli with fMRI. To this end, we examined both intrinsic and dedicated models of time perception by evaluating the tuning properties of sensory and other brain regions given different stimulus durations. We recorded whole brain fMRI signals while participants made bisection judgments of the duration of visual stimuli in the sub-second range. The stimuli were flashed Gabor patterns drifting at different speeds—a manipulation previously shown to modulate their subjective duration. For every voxel, we used training data to estimate the hemodynamic response function and tuning curve to different physical durations. The fMRI response monotonically increases with the stimulus duration mainly in the visual cortices. U-shaped or inverse U-shaped tuning curves are observed in various regions including anterior insula, paracingulate cortex, putamen, anterior cingulate cortex and distributed frontal and parietal regions. For each participant, we then used Bayesian decoder to decode the subjective duration of testing data with 6-fold cross-validation using voxels whose responses were modulated by stimulus durations. Decoded duration correlated well with physical duration when the decoding employed voxels from occipital, temporal or parietal cortex. However, the correlation weakened when decoding from voxels in the frontal cortex. Finally, the correlation disappears when decoding from voxels in basal ganglia or insula, suggesting that these latter regions may encode task difficulty more than duration in particular. In summary, the data suggest both sensory and frontal regions encode subjective duration.
Meeting abstract presented at VSS 2014
This PDF is available to Subscribers Only