Seeing and extrapolating motion trajectories share common informative activation patterns in primary visual cortex

Abstract

The natural environment is dynamic and moving objects become constantly occluded, engaging the brain in a challenging completion process to estimate where and when the object might reappear. Although motion extrapolation is critical in daily life—imagine crossing the street while an approaching car is occluded by a larger standing vehicle—its neural underpinnings are still not well understood. While the engagement of low‐level visual cortex during dynamic occlusion has been postulated, most of the previous group‐level fMRI‐studies failed to find evidence for an involvement of low‐level visual areas during occlusion. In this fMRI‐study, we therefore used individually defined retinotopic maps and multivariate pattern analysis to characterize the neural basis of visible and occluded changes in motion direction in humans. To this end, participants learned velocity‐direction change pairings (slow motion‐upwards; fast motion‐downwards or vice versa) during a training phase without occlusion and judged the change in stimulus direction, based on its velocity, during a following test phase with occlusion. We find that occluded motion direction can be predicted from the activity patterns during visible motion within low‐level visual areas, supporting the notion of a mental representation of motion trajectory in these regions during occlusion.

Contextual modulation of primary visual cortex by temporal predictability during motion extrapolation

“…First, we observed that representational patterns of activity from the visible phase could be used to decode patterns from the occluded phase in upper and lower V1, thereby extending the results from the univariate analyses, where we only found effects in upper V1. Our results suggest that visible and extrapolated types of information share similar representational pattern in the very same voxels, replicating the results from our previous study (Agostino et al 2023). The appropriateness of the sphere selection approach that we used was further confirmed by computing the overlap of sphere centres.…”

“…Reaction time results during visible stimulation showed that participants attempted to estimate the time-to-contact of the stimulus accordingly (faster responses during fast motion, slower responses during slow motion), in HP and LP context, following the expected pattern, and corroborating the results from our first study (Agostino et al, 2023). However, we, further, observed an interaction between velocity, predictability and direction which indicated that participants responded faster to slow downward motion during the LP context.…”

“…In contrast, the other two analyses presented very similar averaged decoding accuracies. Also note that the decoding accuracies of this analysis were generally smaller compared to a previous study (Agostino et al, 2023) as we used trial-wise rather than run-wise estimates as input to the MVPA, which includes an increased noise level, while the previous study used run-wise averages.…”

“…In our previous study (Agostino et al, 2023) we used the 5% highest accuracy approach and showed that the overlap of significant spheres across analyses is above 80% and at the locations of V1 which represented the stimulus trajectory.…”

“…Note that the results from both sessions with opposing velocity association were collapsed, so that a decoding of stimulus velocity cannot explain the results below. Here we expected to decode information from the occluded data with training in the visible phase, as previously observed (Agostino et al, 2023). In order to obtain the most informative voxels, we averaged the 5% most informative spheres (range of 45 to 112 selected spheres).…”