Talis Bachmann

Hearing by Eye: How Much Spatial Degradation can Be Tolerated?

In the McGurk effect (McGurk and MacDonald, 1976 Nature 264 746–748), illusory auditory perception is produced if the visual information from lip movements is discrepant from the auditory information from the voice. A study is reported of the tolerance of the effect to varying levels of spatial degradation (videotaped images of a speakers face were quantised by a mosaic transform). The illusory effect systematically decreased with an increase in the coarseness of the spatial quantisation. However, even with the coarsest level (11.2 pixels/face) the illusion did not completely disappear. In addition, those participants who did not experience the illusion nevertheless showed the effects of auditory–visual interaction in their clarity ratings of the auditory stimulus. It is concluded that auditory – visual interaction in visible speech perception is based on relatively coarse-spatial-scale information.

Integration and Interruption in the Masking of Form by Form

Pairs of geometric forms of equal area were presented, one form after another, with interstimulus intervals t ranging from 0 to 250 ms. The subjects task in experiment 1 was to recognise both stimuli. Identification of the form presented first (backward masking) across all values of t was of a nonmonotonic nature, with greatest impairment at values of t from 30 to 60 ms. Identification of the second form presented (forward masking) increased monotonically with increasing t. Different forms were recognised with different effectiveness across all values of t where masking took place. The results are interpreted on the basis of a multiple-stage processing model which assumes interaction between stimuli at different levels at different values of t. In experiment 2 the subjects task was to detect the presence or absence of a predesignated form. This condition yielded monotonic functions for both backward and forward masking. An explanation in terms of attention-dependent masking is given.

Visibility of Brief Images: The Dual-Process Approach

If successive, brief visual images are exposed for recognition or for psychophysical ratings, various effects and phenomena of fast dynamics of conscious perception such as mutual masking, metacontrast, proactive enhancement of contrast, proactive speed-up of the latency of subjective visual experience, the Fröhlich Effect, the Tandem Effect, attentional facilitation by visuospatial precuing, and some others have been found. The theory proposed to deal with these phenomena proceeds from the assumption that two types of brain processes are necessary in order to consciously recognize visual stimuli: (1) fast, specific processes of encoding that allocate and reactivate the stimulus representation which is based on the activity of selected cortical neurons and (2) relatively slower processes of facilitation of the activity of this specific representation that are mediated by the excitatory modulation of the EPSPs of those selected cortical neurons by the ascending input from nonspecific thalamus. The perceptual retouch construct is proposed in order to characterize and analyze the interaction of (1) and (2). The neurophysiological characteristics of this bifunctional system of afference help to put forward several predictions that are found to be consistent with the empirical regularities of the above-described perceptual-attentional phenomena. These data form a body of converging evidence that is consistent with the predictions of the perceptual retouch approach.

Chapter 8 Twelve spatiotemporal phenomena and one explanation

A metatheory of various “pertentional” effects and phenomena is presented. The concept of pertention is introduced to provide a common denominator for a variety of perceptual and attentional phenomena that may be based on the workings of the same psychophysiological mechanism. Various experimental paradigms that have been used to investigate pertentional events in isolation are brought together on the assumption that a common, two-process (perceptual retouch) theory may be capable of explaining the majority of the effects found. The theory departs from the assumption that most of the data on which our theorising is grounded is itself a function of the subjects reports on the phenomenal content of their perceptions. The subjects who are presented with spatiotemporal physical events in the laboratory are capable of providing this data insofar as conscious representations of the physical events are created. The ample evidence about high-level preconscious processing forces one to assume that the processing which is mediated by the specific data processing cortical modules is not sufficient to guarantee—through the subjects reports—the empirical basis for our theories. This necessarily demands that we look for the neurophysiological mechanisms that are necessary in order to upgrade the specific activity so as to form directly reportable (“conscious”) representations. The characteristics of the socalled nonspecific system of modulation and the operating principles of this system in its interaction with cortical specific modules forms the basis for several experimentally testable predictions about the psychophysical outcomes of pertention when two brief, spatially localised stimuli (stimulus events) are presented to subjects in rapid succession. The spontaneous interaction between the fast specific system of data representation and the slow nonspecific system of modulation of this representation quite naturally explains a wide variety of perceptual and attentional effects and phenomena such as metacontrast, mutual masking, proactive facilitation, paradoxical phenomena related to visible persistence, masking with simultaneous onset/asynchronous offset, line motion illusion, facilitation of visuospatial attention by precues, and several others. The theoretical implications of the present metatheory are discussed and links to other approaches outlined.

The Effects of Coarseness of Quantisation, Exposure Duration, and Selective Spatial Attention on the Perception of Spatially Quantised (‘Blocked’) Visual Images

The two objectives of the present article are (a) to present a brief overview of the effects of systematic variation of the spatial-scale value of quantisation and stimulus duration on the identification of original images that have been degraded by Harmon–Julesz type of image pixelisation by ‘blocking’ and (b) to report the results of two experiments where the effects of selective spatial precuing by local and global peripheral precues on the identification of quantised target stimuli have been studied. Both the overview and the new results reported here demonstrate some counterintuitive effects: (1) abrupt decrease in identification efficiency with only a minor change in the coarseness of quantisation over a critical value of pixels per stimulus; (2) a cost for valid attentional precuing with coarse-quantised images. If physical precues (exposed with stimulus onset asynchrony of 120 ms) were employed in order to orient spatial attention to perceive original or fine-quantised stimuli, then attentional facilitation was found. However, if the precued stimuli were coarse quantised then the facilitative effect crossed over to a detrimental effect of attention. These effects are discussed in the context of the microgenetic approach that presupposes the existence of a perceptual–attentional processing routine that operates according to the coarse-to-fine time-course rule of selective attentional activation of stimulus representations at various spatial scales.

Time-course of Spatial-attentional Focusing in the Case of High Processing Demand on the Peripheral Precue

Subjects had to identify pairs of spatially overlapping or neighbouring letters exposed successively at various locations around the central fixation. The first letter (S1) in a pair acted both as a physical precue to indicate the S2 location and as a stimulus letter to be identified. The second letter (S2) in a pair acted only as a letter to be identified. The SOA was varied between 0 and 320 ms in different experiments. Comparison of the time-course functions of S2 identification in the conditions of (1) single pair presentation and (2) presentation of the letter pairs in the context of distractor letters that accompanied S2 in time, but occupied alternative spatial positions, revealed that S2 processing from the covert spatial-attentional focus was open to non-decreasing interference from the distractor letters up to 240 ms SOAs. An unexpected result consisted in equal level of S2 identification regardless of whether S1 had to be identified or not. This shows that if the peripheral precue has to be pro...