Deeper neural network models better reflect how humans cope with contrast variation in object recognition

Visual inputs are far from ideal in everyday situations such as in the fog where the contrasts of input stimuli are low. However, human perception remains relatively robust to contrast variations. To provide insights about the underlying mechanisms of contrast invariance, we addressed two questions. Do contrast effects disappear along the visual hierarchy? Do later stages of the visual hierarchy contribute to contrast invariance? We ran a behavioral experiment where we manipulated the level of stimulus contrast and the involvement of higher-level visual areas through immediate and delayed backward masking of the stimulus. Backward masking led to significant drop in performance in our visual categorization task, supporting the role of higher-level visual areas in contrast invariance. To obtain mechanistic insights, we ran the same categorization task on three state-of the-art computational models of human vision each with a different depth in visual hierarchy. We found contrast effects all along the visual hierarchy, no matter how far into the hierarchy. Moreover, that final layers of deeper hierarchical models, which had been shown to be best models of final stages of the visual system, coped with contrast effects more effectively. These results suggest that, while contrast effects reach the final stages of the hierarchy, those stages play a significant role in compensating for contrast variations in the visual system.

Distortions of temporal integration and perceived order caused by the interplay between stimulus contrast and duration

Stimulus contrast and duration effects on visual temporal integration and order judgment were examined in a unified paradigm. Stimulus onset asynchrony was governed by the duration of the first stimulus in Experiment 1, and by the interstimulus interval in Experiment 2. In Experiment 1, integration and order uncertainty increased when a low contrast stimulus followed a high contrast stimulus, but only when the second stimulus was 20 or 30ms. At 10ms duration of the second stimulus, integration and uncertainty decreased. Temporal order judgments at all durations of the second stimulus were better for a low contrast stimulus following a high contrast one. By contrast, in Experiment 2, a low contrast stimulus following a high contrast stimulus consistently produced higher integration rates, order uncertainty, and lower order accuracy. Contrast and duration thus interacted, breaking correspondence between integration and order perception. The results are interpreted in a tentative conceptual framework.

Interaction effects of visual stimulus speed and contrast on postural sway

Manipulating the characteristics of visual stimuli that simulate self-motion through the environment can affect the resulting postural sway magnitude. In the present study, we address the question whether varying the contrast and speed of a linear translating dot pattern influences medial–lateral postural sway. In a first experiment, we investigated whether the postural sway magnitude increases with increasing dot speed, as was previously demonstrated for expanding and contracting stimuli. In a second experiment, we also manipulated the contrast of the stimuli. For reasons that high-contrast stimuli can be considered ‘perceptually’ stronger, we expect that higher-contrast stimuli induce more sway than lower-contrast stimuli. The results of the first experiment show that dot speed indeed influences postural sway, although in an unexpected way. For higher speeds, the sway is in the direction of the stimulus motion, yet for lower speeds the sway is in a direction opposite to the stimulus motion. The results of the second experiment show that dot contrast does affect postural sway, but that this depends on the speed of the moving dots. Interestingly, the direction of postural sway induced by a relatively low dot speed (4°/s) depends on dot contrast. Taken together, our results suggest that interactions between the visual, vestibular and proprioceptive system appear to be influenced by an internal representation of the visual stimulus, rather than being influenced by the external visual stimulus characteristics only.

The spatial summation characteristics of three categories of V1 neurons differing in non-classical receptive field modulation properties

The spatial summation of excitation and inhibition determines the final output of neurons in the cat V1. To characterize the spatial extent of the excitatory classical receptive field (CRF) and inhibitory non-classical receptive field (nCRF) areas, we examined the spatial summation properties of 169 neurons in cat V1 at high (20-90%) and low (5-15%) stimulus contrasts. Three categories were classified based on the difference in the contrast dependency of the surround suppression. We discovered that the three categories significantly differed in CRF size, peak firing rate, and the proportion of simple/complex cell number. The classification of simple and complex cells was determined at both high and low contrasts. While the majority of V1 neurons had stable modulation ratios in their responses, 10 cells (6.2%) in our sample crossed the classification boundary under different stimulus contrasts. No significant difference was found in the size of the CRF between simple and complex cells. Further comparisons in each category determined that the CRFs for complex cells were significantly larger than those for simple cells in category type I neurons, with no significant differences between simple and complex cells in category type II and type III neurons. In addition, complex cells have higher peak firing rates than simple cells.