Hans-Christoph Nothdurft

Feature analysis and the role of similarity in preattentive vision

Texture arrays of line elements at various orientations were used to study three phenomena of preattentive vision. Subjects were asked (1) to discriminate texture areas and to distinguish their form (experiments on texture segmentation); (2) to detect salient or vertical line elements (experiments on pop-out); and (3) to identify configurations of similar or dissimilar targets (experiments on grouping). Within the patterns, line orientation was systematically varied to distinguish the effect of differences between areas from the effect of similarity within areas. In all of the experiments, performance was found to depend on local orientation contrast at texture borders rather than on the analysis of line orientation itself. Texture areas were correctly identified only when the orientation contrast at the border well exceeded the overall variation of line orientation in the pattern. Similarly, only target elements with high local orientation contrast were detected fast and “in parallel.” Targets with an orientation contrast lower than background variation required serial search. Preattentive grouping was found to depend on saliency, as defined by local orientation contrast, but not on the similarity of line elements. In addition to local orientation contrast, which played an important role in all of the visual phenomena studied, influences from the alignment of line elements with the outline of a figure were also seen.

Faces and Facial Expressions do not Pop Out

Subjects were asked to detect faces or facial expressions from patterns with a variable number of nonfaces or faces expressing different emotions. In most tests, reaction time was found to increase steeply with sample size, thus indicating serial-search characteristics for the patterns tested. There were, however, considerable differences in the slopes of the graphs (search time versus sample size), which could be attributed to visual (but not face) cues that are discriminated at similar speeds. Slopes did not change when patterns were presented upside down, although such a modification strongly affects the perception of faces and facial expressions.

Neuronal correlates of pop-out in cat striate cortex

Neuronal responses to static and moving texture patterns were investigated in the striate cortex of anaesthetized and paralysed adults cats. Texture patterns were composed of a central light bar presented in the excitatory receptive field of a cell and an array of many similar elements in the surround. For the static condition, elements in the surround were either parallel or orthogonal to the centre line (orientation test). For the moving condition, centre and surround elements (all at same orientation) moved either in the same or in the opposite directions (motion test). Thirty-six percent (31/86) of the neurons tested for motion and 24% (24/99) of the neurons tested for orientation responded more strongly to the patterns displaying feature contrast than to the uniform patterns. These neurons may form a neural basis for visual pop-out of orientation and motion.

Orientation sensitivity and texture segmentation in patterns with different line orientation

Orientation sensitivity (OS) and the ability of human subjects to discriminate structured areas of different texture orientation (DOT) were investigated using line arrays of varying line length. In general, OS was mediated by shorter lines than was DOT; lines can be distinguished by their orientation before they give the impression of a texture border between adjacent regions with lines differing in orientation. This difference was found to hold over a range of retinal eccentricities from 5 degrees nasal to 30 degrees temporal. Decreasing visual acuity, associated with increasing distance from the fovea, cannot, however, account for the higher threshold of DOT, even taking into account that DOT requires a larger area for analysis than OS. When angle of orientation between adjacent texture areas was varied instead of line length, DOT thresholds at different retinal locations were reached at similar values. The difference between OS and DOT, found consistently at all retinal positions, suggests that they are mediated by distinct neural mechanisms.

Attention shifts to salient targets

To investigate the role of salience in fast visual search, time courses for the detection and identification of salient targets were measured in six subjects, using texture-like line arrays. Single lines were made salient from luminance contrast, motion contrast, or by an added circular cue that is known to attract focal attention. Three major findings are reported: (1) Identification of target orientation required longer presentations than detection of the saliency effect itself, consistent with the model that target salience attracts focal attention for target analysis. (2) Different saliency mechanisms produced similar effects, suggesting that salience from feature contrast is functionally equivalent to salience evoked from a visual cue. (3) Circular cues were most effective when presented close to the target; performance in target identification decreased when the diameter was enlarged so that the cue was presented farther away and on a different spatial scale. All together, these findings suggest that popout targets in visual search may be detected fast and independent of set size because (a) they are salient and attract focal attention, and (b) their salience is produced on the same spatial scale and at the same location in the visual field where target properties are encoded.

Evidence for asynchronous development of sleep in cortical areas

WE have recorded from extrastriate area V4 in monkeys performing a visual search task. When animals became tired or drowsy, responses to visual stimulation were often reduced or even completely blocked, and background activity changed to the burst-pause pattern typically seen in sleep. In spite of such neuronal sleep observed in V4, animals continued to perform the visual task, indicating that at least the primary visual cortex was still working. This observation shows that sleep does not develop simultaneously in all cortical areas but may affect some areas earlier than others. In particular conditions, local sleep of certain areas may be a stable and long-lasting phenomenon.

Salience from feature contrast: variations with texture density☆

The salience of popout targets was measured in regular line arrays as a function of texture density. Test targets (singletons with orientation, motion, or luminance contrast) presented at different raster widths were compared with reference lines (lines brighter than surrounding lines) presented at fixed raster width. The luminance at which the reference target appeared as salient as the particular test target was taken as a measure of the relative salience of the test target. For orientation or motion contrast, targets at medium to small raster widths were far more salient than targets in sparse or very dense line arrangements. For targets defined by luminance contrast, salience variations with texture density were less pronounced. Some subjects also reported salience for lines in sparse arrangements even when these did not display feature contrast. When such non-specific saliency effects were subtracted from the actual measurements, salience curves for orientation or motion contrast revealed peaks of increased sensitivity at line spacings below 2-3 deg and flat curves at larger grid sizes. In an additional experiment, saliency effects from orientation contrast were measured using texture lines of different size. Salience variations were commonly observed. However, the curves were not found to scale with the different sizes of texture elements but were constantly related to the free space between neighbouring lines. This suggests that peaks in the salience profiles reflect the limited spatial extent of the underlying neural mechanisms.

Neuronal responses to orientation and motion contrast in cat striate cortex

Responses of striate neurons to line textures were investigated in anesthetized and paralyzed adult cats. Light bars centered over the excitatory receptive field (RF) were presented with different texture surrounds composed of many similar bars. In two test series, responses of 169 neurons to textures with orientation contrast (surrounding bars orthogonal to the center bar) or motion contrast (surrounding bars moving opposite to the center bar) were compared to the responses to the corresponding uniform texture conditions (all lines parallel, coherent motion) and to the center bar alone. In the majority of neurons center bar responses were suppressed by the texture surrounds. Two main effects were found. Some neurons were generally suppressed by either texture surround. Other neurons were less suppressed by texture displaying orientation or motion (i.e. feature) contrast than by the respective uniform texture, so that their responses to orientation or motion contrast appeared to be relatively enhanced (preference for feature contrast). General suppression was obtained in 33% of neurons tested for orientation and in 19% of neurons tested for motion. Preference for orientation or motion contrast was obtained in 22% and 34% of the neurons, respectively, and was also seen in the mean response of the population. One hundred nineteen neurons were studied in both orientation and motion tests. General suppression was correlated across the orientation and motion dimension, but not preference for feature contrast. We also distinguished modulatory effects from end-zones and flanks using butterfly-configured texture patterns. Both regions contributed to the generally suppressive effects. Preference for orientation or motion contrast was not generated from either end-zones or flanks exclusively. Neurons with preference for feature contrast may form the physiological basis of the perceptual saliency of pop-out elements in line textures. If so, pop-out of motion and pop-out of orientation would be encoded in different pools of neurons at the level of striate cortex.

The conspicuousness of orientation and motion contrast.

Subjects were asked to compare the salience of different targets in a texture array. Elements that popped out from local differences in orientation or motion were compared with texture elements that popped out from luminance because they were brighter than all other elements in the pattern. The luminance of a comparison target that produced an equal preference rating for both targets was taken as a direct measure of the salience of the target under test. Saliency matches were made for targets in different background conditions; orientation and motion defined targets were tested in separate experiments. The data show that the salience of pop-out targets does not linearly increase with increasing orientation or motion contrast but reveals nonlinear properties from threshold and saturation effects. With increasing variation of background elements, target salience decreases continuously. Already small variations in background texture make a given target appear less compelling than when presented on a homogeneous background. The data also show that orientation and motion properties behave very similarly in these aspects and suggest that a similar neuronal mechanism may underly pop-out in both visual dimensions.

Focal attention in visual search

Visual search operates in different modes assumed to reflect serial and parallel processing. The basis of this distinction is not yet clear. It is often assumed that serial search involves sequential shifts of focal attention across a scene and that no such shifts occur in parallel search. Direct measurements of attention effects during search show that the focus of attention moves to the target (and away from non-targets) both in serial and parallel search. This suggests that the two search modes do not differ in their attentional load but perhaps in the way in which focal attention is directed to the target.