Purchase this article with an account.
Naotsugu Tsuchiya, Alexander Maier, Nikos Logothetis, David Leopold; Neuronal activity in area MT during perceptual stabilization of ambiguous structure-from-motion. Journal of Vision 2009;9(8):756. doi: 10.1167/9.8.756.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
During continuous viewing of an ambiguous display, perception spontaneously alternates between two alternative interpretations of the stimulus. Such perceptual alternation can be greatly slowed down all the way to a standstill if the stimulus is periodically removed from view, a phenomenon called ‘perceptual stabilization’ (Leopold et al NatNeuro 2002). While brain activity during perceptual switches has been studied extensively, neuronal activity during perceptual stabilization remains unexplored. Here, we report neuronal responses in area MT in two behaving macaque monkeys under two conditions that differ in the strength of perceptual stabilization. In the first condition, we presented ambiguous structure-from-motion (SFM) stimuli periodically, a standard paradigm for stabilization in humans. The monkeys' behavioral reports confirmed minute-long stabilization and their ocular responses mirrored the perceptual reports. In the second condition, we intermixed ambiguous SFM with disparity-defined unambiguous SFM across trials. In this latter condition, the monkeys did not show any evidence for perceptual stabilization. We analyzed neuronal activity that was simultaneously recorded with ∼10 electrodes in the form of multi-unit spiking activity and local field potentials using multivariate decoding techniques (Tsuchiya et al PLoSONE 2008). Using linear decoders, we were able to predict the percept during stimulus presentation (∼70% correct), but not during the pre-stimulus blank periods in the stabilization condition. A significant difference between the stabilization and the non-stabilized conditions was the latency of decoding accuracy; while the decoding accuracy reached its peak within 0.2–0.3 sec from the stimulus onset in the stabilization condition, it reached the peak at ∼0.7 sec in the non-stabilization condition. We conclude that while MT does not represent explicit information during pre-stimulus blank periods, it is highly primed in the stabilization condition such that upon stimulus onset its activity converges towards one perceptual state much faster than in the non-stabilized condition.
This PDF is available to Subscribers Only