Perception

The characteristics of human sniffing episodes during odour perception have been described in an earlier paper, where it has been suggested that the techniques used by individuals may be close to those providing optimum perception. To investigate this suggestion, threshold and intensity tests with butanol, cyclohexanone, and pentyl acetate have been carried out on twenty-one subjects. Olfactory responses obtained by using natural sniffing techniques were compared with those where the number of sniffs, interval between sniffs, and size of sniffs were varied. The results indicate that it is very difficult to improve on the efficiency of sniffing techniques of individuals and that a single natural sniff provides as much information about the presence and intensity of an odour as do seven or more sniffs. A single natural sniff and the first sniff of a a natural sniffing episode were shown to have similar characteristics and most significantly both were unaffected by changes in the concentration and type of odour. Overall, the results indicate that humans achieve optimum odour perception during threshold and intensity measures with their natural multiple-sniff technique or with a single sniff. For the ‘average’ human this occurs with a sniff of inhalation rate 30 1 min⁻¹, volume 200 cm³, and duration 0.4 s. The use of several sniffs in a sniffing episode appears to be a confirmatory action rather than a necessary one, except for the perception of odour mixtures where several sniffs are likely to be needed to aid discrimination of the components. Data from the present and earlier study provide the information necessary for the development of a standard olfactometer and standard procedures for measuring the olfactory responses of humans.

Correlational methods were used to investigate symmetry of effect for the arrowhead and featherhead versions of the Müller-Lyer figure. Two control figures were compared in the determination of baseline levels for measurement of the illusions: a shaft presented without any inducing context, and a shaft with vertical inducing lines attached. In addition, results based on difference-score measures of the illusions were contrasted with results obtained by partial-correlation techniques. Overall, when one considers the results for either one of the arrowhead or featherhead versions, the evidence favours a common underlying mechanism. However, results across the two versions suggest that the mechanisms for the two versions differ fundamentally. In weighing the different kinds of evidence contributing to this conclusion, methodological issues were raised. By obtaining two judgments for each stimulus figure from a large number of subjects, it was possible to demonstrate not only that conventional difference-score measures of illusions are highly unreliable, but also that they can yield biased results.

The Café Wall illusion is a distortion illusion in which the parallel lines of a chessboard-like figure consisting solely of parallel and perpendicular line elements appear to converge in alternating rows, creating a wedge distortion similar to that of the well-known Zöllner illusion. Gregory and Heard have formulated an explanation for the Café Wall illusion which relies upon the operation of a ‘border-locking mechanism’ in the visual system. The results of the present experiment suggest an alternative explanation in which the operation of brightness induction within the mortar regions of the Café Wall produces a series of ‘twisted cords' or slanted line elements akin to those of the Fraser or Zöllner figures. A series of such ‘twisted cords' is shown to be capable of itself to produce an illusory convergence like that of the Café Wall. Manipulations of the luminance of discrete regions in the mortar lines of the Café Wall, designed either to augment or cancel the effects of brightness induction in the production of these slanted line elements, are successful in enhancing or reducing, respectively, the wedge distortion of this visual illusion.

Radial lines pointing towards a central circular region induce an illusory positive brightness contrast. Shift in line orientation from radial to tangential can result in negative brightness contrast and line enhancement.

The viewing system parameters (in particular the mutual orientation of the coordinate frames fixed respectively to the left and right projection surface) probably have to be explicitly incorporated into solutions to the stereoscopic-matching problem in binocular vision. An algorithm for computing the relative orientation of the two directions of gaze from purely visual information without a prior solution to the correspondence problem is outlined. The availability of an algorithm of this sort seems to be a precondition for any stereoscopic-matching process that uses the epipolar constraint. It is further argued that fixation of objects in the near space requires integration of the results of stereoscopic matching over eye movements, and that this requirement puts rather tight restrictions on the possible form of representing the results of stereoscopic processing.

The horizontal extent of Panum's fusional area was measured by means of a single-vertical-line stimulus placed at thirty-two locations throughout the peripheral visual field. These results were transformed by using known values of the human cortical magnification factor (CMF), and the hypothesis that variations in the magnitude of Panum's area may be accounted for by variations in the CMF was tested. It was found that the increase in Panum's area with increasing stimulus eccentricity correlates well with the CMF, but that variations in the extent of Panum's area as a function of angular position around the line of sight do not correspond well with the CMF.

The effects of brightness, hue, and saturation on perceived depth between adjacent regions have been examined. The stimulus consisted of two hemifields of different colors, and the subject was asked to state which appeared nearer and to judge the perceived depth between them. When both hemifields were achromatic, the perceived depth was found to increase with increasing brightness difference. Some subjects tended to judge the brighter side nearer, others the darker side nearer. With the achromatic–chromatic combination, there were no differences in perceived depth among three hue conditions, whilst with the chromatic–chromatic combination the perceived depth depended on hue combination. In terms of decreasing frequency of ‘nearer’ judgments the hue order was red, green, blue. When the two hemifields differed only in saturation, the perceived depth increased with increasing saturation difference, and whether the more saturated or the less saturated side was judged nearer depended on hue. It is argued that the effects of brightness and saturation on perceived distance from the observer can be attributed to figure–ground differentiation between adjacent regions in the visual field; but this argument does not cover the effect of hue under achromatic background conditions.

The aim of the experiment was to study the evolution with age (6, 8, 11 and 14 years) of pictorial depth perception in Pandora's box and to compare it with the evolution of size illusion with the same subjects and the same pictorial backgrounds. In addition to familiar size and relative position, each pictorial stimulus contained one or more of the following depth cues: linear perspective, texture gradient, and interposition. The two kinds of measurements produced different results. Size illusions, although present, did not vary with age but increased with the number of cues. Estimates of distance in Pandora's box increased with age and varied according to the type of cue present: texture gradient seemed to be critical to the amount of depth perceived. The correlation between size adjustments and distance adjustments was significant only for the two oldest groups of subjects (11 and 14 years).

It is well-known that patterns of eccentric circles when slowly rotated give rise to compelling three-dimensional impressions of cones or conical holes which can ‘wobble’ as the pattern rotates. The wobble can be considered as part of the overall phenomenon of depth elicited from a rotating display, the ‘stereokinetic’ effect (SKE). This paper considers the three-dimensional appearance as being the result of the sliding of contours and thus it imitates the motion parallax found in real three-dimensional objects in motion. New variants of SK figures are used to examine these points. An analogy with computer programs is proposed which questions earlier views on the location of perceptual invariance.

The relevance of low and high spatial-frequency information for the recognition of photographs of faces has been investigated by testing recognition of faces that have been either low-pass (LP) or high-pass (HP) filtered in the spatial-frequency domain. The highest resolvable spatial frequency was set at 15 cycles per face width (cycles fw⁻¹). Recognition was much less accurate for images that contained only the low spatial frequencies (up to 5 cycles fw⁻¹) than for images that contained only spatial frequencies higher than 5 cycles fw⁻¹. For faces HP filtered above 8 cycles fw⁻¹, recognition was almost as accurate as for faces LP filtered below 8 cycles fw⁻¹, although the energy content of the latter greatly exceeded that of the former. These findings show that information conveyed by the higher spatial frequencies is not redundant. Rather, it is sufficient by itself to ensure recognition.

A critical difficulty in theories of the ‘mental rotation’ phenomenon has been to find a computationally plausible reason why the rotation should occur in small intermediate steps. It is pointed out that this difficulty is peculiar to metrical representations: if spatial relations are presented symbolically but nonmetrically, then the iterative or recursive application of minimal transformations is memory saving. A program rotter is described to illustrate this principle.

What effect may an unattended word have during a single fixation? Attention was selectively directed to a word exposed for 50 ms by the demand to make a rapid lexical decision response, and during the same presentation a second word was displayed approximately 2.3 deg of visual angle away on the same horizontal axis. The second word was backwards and forwards masked by a random-dot display, and was described to subjects as a distractor which was to be ignored. Although no response was required to this word, it was found to influence response latency to the attended word according to the semantic relationship which existed between the two. They interfered with the lexical decision response when the two words were related in meaning and also when they sounded as if they were related in meaning. These effects argue for automatic processing of the meaning of printed words presented in either visual hemifield, and for their automatic phonological activation.

A research programme has been carried out that concerns the accuracy with which listeners can identify a speaker heard once before. The present study examined the voice-recognition abilities of blind listeners, and it was found that they could more accurately select target voices from the test arrays than could sighted people. However, the degree of blindness, the age at onset of blindness, and the number of years of blindness all failed to relate to voice-recognition accuracy.