S. Mateeff

Dynamic Visual Capture: Apparent Auditory Motion Induced by a Moving Visual Target

Apparent motion of a sound source can be induced by a moving visual target. The direction of the perceived motion of the sound source is the same as that of the visual target, but the subjective velocity of the sound source is 25–50% of that of the visual target measured under the same conditions. Eye tracking of the light target tends to enhance the apparent motion of the sound, but is not a prerequisite for its occurrence. The findings are discussed in connection with the ‘visual capture’ or ‘ventriloquism’ effect.

Selective directional sensitivity in visual motion perception

We present two experiments demonstrating that: (i) the latency of perception of the position of a small visual target moving towards the fovea is shorter than that of the same target moving away from the fovea; (ii) the reaction time (RT) to onset of motion of the same type of target is also shorter when it moves towards the fovea; and (iii) the RT to onset of motion away from the fovea may be shorter when larger, textured stimuli are employed. The relation of the findings to the existence of two systems for visual motion information processing and to recent neurophysiological findings is discussed.

The role of the adjacency between background cues and objects in visual localization during ocular pursuit

Subjects used eye movements to pursue a light target that moved from left to right with a velocity of 15 deg s−1. The stimulus was a sudden five-fold decrease in target intensity during the movement. The subjects task was to localize the stimulus relative to either a single stationary background point or the midpoint between two points (28 deg apart) placed 0.5 deg above the target path. The stimulus was usually mislocated in the direction of eye movement; the mislocation was affected by the spatial adjacency between background and stimulus. When an auditory, rather than a visual, stimulus was presented during tracking, target position at the time of stimulus presentation was visually mislocated in the direction opposite to that of eye movement. The effect of adjacency between background and target remained the same. The involvement of processes of subject-relative and object-relative visual perception is discussed.

The time it takes to detect changes in speed and direction of visual motion

We studied the ability of human observers to detect abrupt changes in velocity of motion of a random dot pattern. The pattern moved horizontally for 0.9 s at velocity V0, then changed to V1 either in speed, or in direction for a time T and returned to the initial motion. The threshold duration for detection of the change was measured for initial speeds of 2, 4, 8 and 16 deg/s. The time to detect a velocity reversal was equal to that for detection of an increase in speed by a factor of three. The time to detect an abrupt cessation of motion was equal to the time for detection of an increase in speed by a factor of two. The time to detect a direction change, the speed being constant, decreased gradually with increasing angle between V0 and V1 from 12 to 180 degrees and with increasing V0; the detection time was a function of (V1-V0) almost independent of the value of V0. This finding supports the hypothesis of Dzhafarov et al. (Percept Psychophys 1993;54:373-750), that the visual system effectively reduces the detection of velocity changes (from V0 to V1) to the presumably more simple detection of a motion onset, from 0 to (V1-V0). The characteristics of the detection process in the cases of uni- and two-dimensional velocity changes are discussed.

Temporal thresholds and reaction time to changes in velocity of visual motion

A random dot pattern moved at a velocity V1. The velocity then increased or decreased abruptly to another value V2 for some time and again returned to V1. The temporal threshold, i.e. the duration of V2 that was necessary to detect the change was measured. Thresholds for the detection of the same velocity increment, V2 = 2 x V1, were shorter when the baseline velocity V1 increased from 1 to 8 deg/sec (Expt 1). The temporal threshold decreased as the velocity contrast (V2 - V1)/(V1 + V2) increased from 0.33 to 0.77. The thresholds for the detection of velocity decrements were in general longer than those for the detection of increments (Expt 3). In Expts 2 and 4 the random-dot pattern moved with velocity V1, which abruptly increased or decreased to V2, without returning to V1. The reaction time to the change was measured for the same velocity pairs as those used in the temporal threshold measurements. There was a good correspondence between changes in the reaction times and changes in the thresholds under the various conditions. The data are interpreted on the basis of two hypotheses: higher velocities are detected by mechanisms that respond more rapidly; and integration of velocities occurs when temporally-adjacent motions are presented.

The relation between the velocity of visual motion and the reaction time to motion onset and offset

Responding to visual motion is a basic function of the nervous system of most animals. It is therefore curious that this function, and in particular the relationship between reaction time (RT) and stimulus velocity, has rarely been investigated. There are few studies concerned with the human ability to react as quickly as possible after an object (or texture) starts to move at a constant velocity in the observer’s visual field (Mashhour, 1964; Ball & Sekuler, 1980; Tynan & Sekuler, 1982). It has been established that the RT to motion onset decreases with increasing velocity of the stimulus. Mashhour (1964) found that the relationship between RT and velocity can be approximated by a formula of the type

Localization of brief visual stimuli during pursuit eye movements

Experimental findings concerning the properties of the phenomenon of mislocation of brief visual stimuli during smooth eye tracking are described. One of these, which cannot be explained by existing hypotheses, is that under certain conditions the mislocation magnitude tends to have zero or even negative values. A model is developed for explanation of the mislocation phenomenon. It is suggested that localization is based on: (1) information about the current eye position and (2) information about the stimulus locus on the retina. They both arrive at the localization centre with non-zero delays. The mode of information processing in this centre leads to a magnitude of mislocation which is proportional to the difference between the two delays and which could be positive, zero or negative. Factors which influence either delay should also influence the mislocation magnitude.

Motion Extrapolation Performance: A Linear Model Approach:

Researchers have obtained similar results from different visual motion extrapolation experiments despite the large variety of motion stimuli used. With respect to the ability of human subjects to judge the moment at which an occluded moving stimulus arrives at a predetermined position along its motion path, the general conclusion has been that errors increase with the duration of the occluded motion. However, substantial individual differences are often obscured within this statement. We propose a linear model to describe the performance of human observers in motion extrapolation tasks. The results from an experiment on centrifugal and centripetal motion extrapolation are examined in terms of this model. We discuss the restrictions imposed by the model on conclusions drawn after converting estimated arrival times to velocity estimates or accuracy scores. The parameters of the linear regression describing the individual performance in motion extrapolation tasks might be appropriate measures of interindividual differences.

A constant latency difference determines directional anisotropy in visual motion perception.

In a recent paper in this journal (Mateeff, Yakimoff, Hohnsbein, Ehrenstein, Bohdanecky & Radil, 1991) we reported a new visual anisotropy: inward (foveopetal) motion of a small visual target was perceived with a latency shorter than the perceptual latency of outward ( foveofugal) motion. The latency difference was demonstrated by a simultaneity judgement task. A visual target moved horizontally toward or away from a fixation point. Another stationary target was positioned just above the path of the moving target at a distance of 16 deg from the fixation. A fovea1 light signal was given during the motion and the observer had to judge whether the moving and stationary peripheral target were aligned at the moment the signal occurred. Under the condition of inward motion apparent alignment occurred more than 100 msec earlier than in the case of outward motion. The possibility that the different performance under the two directions of motion might be due to less accurate localization of targets in the visual periphery was ruled out by the results of a control task. In this task the moving target disappeared and the observers had to localize the place of disappearance relative to the stationary target. Although their accuracy in determining the position of the moving target at the moment of its disappearance was not perfect, the mislocations obtained could not explain quantitatively the data from the simultaneity judgement task. It was concluded that centripetal motion of visual targets is indeed perceived faster than centrifugal motion, There is another way to distinguish the effects of spatial and temporal factors in the localization of moving targets: by measuring the moment of subjective alignment between the stationary and the moving targets for different target velocities. If performance is determined only by a constant perceptual delay, the latency difference between inward and outward motion should remain the same as velocity is varied. Correspondingly, the localization errors should increase with increasing velocity. On the other hand, if performance is determined only by a constant spatial error, then the mislocation should remain the same at different velocities, implying that the latency difference between inward and outward motion diminishes with increasing velocity. In the two experiments reported here the main and the control tasks of the first experiment described by Mateeff et al. (1991) were replicated for different motion velocities.

The simple reaction time to changes in direction of visual motion.

Abstract Recently Dzhafarov et al. presented a model explaining data on simple reaction time (RT) to unidimensional velocity changes. The authors suggested that having a motion with an initial velocity V0, the velocity change detection system is reinitialized by means of a ”subtractive normalization” process. Therefore, any abrupt change from V0 to V1 is detected as if it were the onset of motion with a speed equal to |V1–V0|. They derived that the RT is a function of |V1-V0|–2/3. We tested this model for the case of two-dimensional velocity changes. Our subjects observed a random dot pattern that moved horizontally, then changed the direction of motion by an angle α in the range between 6° and 180° without changing the speed V. Speeds of 4 and 12 deg/s were used. The subjects reacted as quickly as possible to the direction change. The RTs asymptotically decreased with increasing α; with 12 deg/s speed the RTs were shorter than those obtained with 4 deg/s. It was shown that the data can be well described as a function of |V1–V0|–2/3=(2Vsin(α/2))–2/3. An extension of the ”subtractive normalization” hypothesis for the case of two-dimensional velocity changes is proposed. It is based on the assumption that the velocity vector V1 after the change is decomposed into two orthogonal components. Alternative explanations based on the use of position or orientation cues are shown to contradict the data.