Eric Chiu

A Test of the Sensorimotor Account of Vision and Visual Perception

Two theories define the relationship between sensory experience and perception of location. The doctrine of specific nerve energies relies on hard-wired, genetically specified relationships between stimulation and perception, modifiable only within limits by adaptation. In a newer sensorimotor account, experience tunes the relationship between stimulation and perception. The perception of pressure phosphenes can differentiate the two theories, because the phosphene appears at a location predicted by physiological optics and in a modality predicted by specific nerve energies. Moving a finger vertically along the outer orbit of the eye while pressing gently on it through the lid during nasally directed gaze results in apparent motion of the phosphene out of phase with the finger, therefore in contradiction to information from motor efference to the finger, tactile sense at the fingertip, eyelid and bulb, joint receptors, and proprioception from muscles driving the finger. A test of the sensorimotor theory giving it every advantage had six observers in darkness moving their fingers along the eye and observing phosphenes for 1 h and 2400 motion cycles; the phosphene always obeyed the doctrine of specific nerve energies, never adapting or changing modality as the sensorimotor theory predicts.

A Slippery Slope: Estimated Slant of Hills Increases with Distance

The slopes of hills tend to be greatly overestimated. Previous studies have found that slope estimates are significantly greater when estimated verbally than with a proprioceptive measure. It has yet to be determined whether these estimates are made for the entire extent of the slope, or whether the estimates in closest proximity are estimated using a different process. Since some parietal cortex neurons respond differently to objects within arms reach, short-distance slope estimation may utilize these or analogous neurons. Alternatively, greater implied effort might make longer slopes seem steeper. We determined that both verbal and proprioceptive reports of slope are overestimates that increase logarithmically with distance from the observer, contradicting both theories. Consistent with previous work, proprioceptive estimates were more accurate at all ranges. Our results can be interpreted as a function of the angle between the observers gaze and the plane of the hill, modified by depth cues available at only near distances.

Timing of Speech and Display Affects the Linguistic Mediation of Visual Search

Recent studies have shown that, instead, of a dichotomy between parallel and serial search strategies, in many instances we see a combination of both search strategies utilized. Consequently, computational models and theoretical accounts of visual search processing have evolved from traditional serial-parallel descriptions to a continuum from ‘efficient’ to ‘inefficient’ search. One of the findings, consistent with this blurring of the serial–parallel distinction, is that concurrent spoken linguistic input influences the efficiency of visual search. In our first experiment we replicate those findings using a between-subjects design. Next, we utilize a localist attractor network to simulate the results from the first experiment, and then employ the network to make quantitative predictions about the influence of subtle timing differences of real-time language processing on visual search. These model predictions are then tested and confirmed in our second experiment. The results provide further evidence toward understanding linguistically mediated influences on real-time visual search processing and support an interactive processing account of visual search and language comprehension.