No description available.

No description available.

When identifying two masked targets presented in rapid succession, awareness of the second may be reduced when it is presented between 100 and 400 ms after the first. This phenomenon has been termed the attentional blink (AB). A wealth of knowledge has been collected regarding performance when both targets are presented visually however, evidence concerning an auditory analogue has been scarce. Nine experiments presented here demonstrate that the auditory attentional blink (AAB) shares some commonalities with but also has some differences from the visual attentional blink (VAB). Two experiments examined cross-modal dual-task interactions and provide only equivocal evidence for a cross-modal AB. All eleven experiments demonstrated the influence of non-target (distractor) items upon target detection. It was shown that presenting targets within an ordered distractor sequence was an important pre-requisite for the AAB. In addition, the level of exposure to the distractor sequence before the presentation of the first target (T1) moderated target identification. Increasing practice (incorporating target and distractors) also attenuates the magnitude of the AAB. In a similar vein, targets of a different stimulus set to that of the distractors also attenuate the AAB. Unlike the VAB, introducing a switch in stimulus set between targets increased performance at early SOAs. For the VAB, very little consideration has been given to items occurring before T1, and the pre-eminent masking role of the +1 item is reflected in all theoretical explanations of the VAB. However, the AAB may rely on items occurring before as well as after the targets. It is well established that the nature of the auditory scene provided by the distractors may change the way that targets are defined and processed. Thus, processing restrictions demonstrated by the AAB may not arise specifically from masking but due to the demands of target extraction from ordered perceptual streams.

The six experiments reported in this thesis tested Endsley's (1995) three level theory of perception, comprehension and projection in SA alongside relevant cognitive theories, using a driving hazard perception test (HPT) of both hazard recognition and hazard anticipation as the main dependent measures. Experiments 1 and 2 tested the effects of training and expertise on the HPT, revealing a positive association between hazard anticipation and both SA training and expertise. Experiment 2 revealed a potentially complex relationship between expertise and recognition and anticipation. The effects of training Endsley's (1995) perception and comprehension levels of SA did not appear to be additive as predicted by that theory. Experiments 3 and 4 tested the effects of concurrent freeze probes and real time SA probes on the HPT. This revealed a negative effect of interruption and a positive effect of reorientation (revealing the rationale of the task) associated with the use of a freeze probe, in terms of hazard anticipation. Furthermore, these two effects appear to be mutually cancelling. It was also found that notification of forthcoming online probes does not ameliorate the negative effects of those interruptions in terms of hazard anticipation. Experiments 5 and 6 tested the effects of working memory interference in terms of visual-spatial, phonological and episodic buffer processing on the two HPT measures. This revealed more deleterious effects on hazard recognition associated with visual-spatial and episodic buffer interference than with phonological interference. It is argued that in terms of SA related visual processing during driving Milner and Goodale's (1992) dual pathway theory appears to have more explanatory power than Endsley's (1995) theory of SA.

Unsupervised exposure to confusable stimuli facilitates later discrimination between them. It is known that the schedule of exposure is critical to this perceptual learning effect, but several issues remain unresolved: I) it is not known whether a mechanism of mutual inhibition, taken by some to underpin perceptual learning in rats, is also evident in humans. II) Although simultaneous presentation of the to-be- discriminated stimuli has been suggested by some to be the most efficient way to promote perceptual learning, the associative mechanisms proposed by others (e.g., that of mutual inhibition) predict the opposite. Ill) Perceptual learning has been invoked as the process by which a face becomes familiar but surprisingly, this idea has received little empirical evaluation. The experimental work reported in this thesis addresses these three issues. Experiments 1 and 2, using flavours as stimuli, reveal that the inhibitory mechanisms that contribute to perceptual learning in rats also contribute to perceptual learning in humans. Experiments 3 and 4 demonstrate a perceptual learning effect using visual stimuli, pictures of human faces and that these effects too, exhibit parallels with studies of perceptual learning with rats. In particular they demonstrate that intermixed exposure results in greater perceptual learning than does blocked exposure. Experiments 5 to 7 indicate that perceptual learning seen following simultaneous exposure is, in turn, superior to intermixed exposure - implicating a process of stimulus comparison. Experiment 8 confirms that this novel effect is also observed with other visual stimuli, chequerboards, while those of Experiments 9 and 10 indicate that the face stimuli used exhibit some of the hallmarks of face processing. These findings establish, along with Experiments 3 to 6, that perceptual learning contributes to the process by which a face becomes familiar.

Two paradigms are presented to investigate the facilitatory effect of spatial cueing on perception of the features of objects. In Chapter 2, performance was studied for continuous tracking of the positional and non-positional features of multiple objects. Observers attempted to report the last feature value of a queried object. The results demonstrated a progressive decline in the precision of the representation of tracked objects, for orientation, spatial period and position. The decline was apparent as the number of objects was increased from one to two to four, and occurred before previously suggested capacity limits based on a fixed number of objects. Responses were more similar to past states of the queried object than to its last state, that is to say that responses exhibited perceptual lags. For orientation, and particularly for spatial period, perceptual lags increased as more objects were added to the attentional load. Analyses of error patterns broken down by likely confidence in responses suggested a contribution from serial processing. Further support for the notion that the lags were contributed to by serial processing, was suggested by two double report experiments presented in Chapter 3. Here, for spatial period tracking, processing appeared not to proceed purely in parallel. Why tracking performance may differ between features is discussed. Four experiments in Chapter 4 rule out the possibilities that the differences in lags between orientation and spatial period were due to artefacts relating to the range of feature values. Chapter 6 is an investigation into the benefits of attention in terms of noise reduction and sampling. Pre-cueing and dual task manipulations were used to compare the effects of attention with those of crowding, which has been demonstrated to increase internal noise. Spatially directed attention was found to reduce internal noise, consistent with the precision results of the previous chapters.

The retrieval processes supporting recognition memory for faces were investigated using event-related potentials (ERPs) and behavioural measures. The ERP old/new effects elicited by faces were investigated in five experiments in which participants were required to distinguish between old and new (studied and non-studied) faces. A direct comparison between the ERP old/new effects elicited by faces and words in an old/new recognition memory task in Experiment 1 provided evidence for at least one common old/new effect, as well as evidence for a material-specific retrieval effect that was only present for faces. The subsequent experiments employed "recognition confidence judgments" (Experiments 2 and 3) and "source memory" manipulations (Experiments 4 and 5) to separate neural activity that might be tied to the processes of recollection and familiarity. Across the two recognition confidence experiments, reliable old/new effects were evident mainly for responses that attracted high confidence judgments, and there was little evidence for modulations that were sensitive to the level of recognition confidence systematically. These data indicate that ERPs index memory processes supporting face judgments that are linked to recollection. The two source memory experiments also revealed superior old/new effects which covered both frontal and parietal scalps and which were larger for those correct old responses that attracted correct rather than incorrect source judgments. The ERP data thus provides strong evidence for neural indices of recollection across all experiments. It might be regarded as surprising that, given the findings in ERP studies with verbal materials, no strong evidence for an ERP correlate of familiarity was found in the ERP data. In Experiment 4, a mid-frontal old/new effect in the 300-500ms time window was present for all correct old responses, and was insensitive to the source judgments, suggesting that this modulation is a neural index of familiarity. This pattern of data, however, was not replicated in Experiment 5 when a more rigorous separation between familiarity- and recollection-based responding was employed. These ERP findings are considered in the context of dual-process theories of recognition memory and their broad application across markedly different kinds of studied materials.

Inversion has been shown to disrupt face recognition, but relatively few studies have looked at the processes involved in the recognition of faces seen at intermediate angles of rotation. Here we address this issue by looking at the processing of rotated faces in a recognition memory task thought to encourage holistic processing, and in the matching of Thatcherised faces, a task which is seen as an indicator of configural processing. When faces were equally rotated at learning and test, we found no evidence of holistic processing. When the task was to recognise differently oriented faces, however, performance declined as an approximately linear function of the difference in orientation between the learning and test faces, suggesting that participants may be mentally rotating faces prior to recognition. Chapter 4 considered the effects of rotation on a same-different matching task thought to encourage configural processing. When the task required the matching of local configural information, the effect of rotation was approximately equal for normal and Thatcherised faces, but Thatcherisation disproportionately disrupted global configural information. The effect of rotation on these forms of information when face pairs contained identical images of the same person, and in an identity-matching task, was also explored. Chapter 6 looks at the effects of inversion on the detection of configural and featural changes to faces in a visual search task. Similar effects of inversion and search strategies were observed for both types of change at both angles of orientation, suggesting that face processing mechanisms do not extract configural at the expense of featural information from faces in this task. The implications of these findings for theories of face processing and the nature of the relationship between rotation and face processing are discussed, and the extent to which the mental rotation hypothesis can account for these findings is also considered.

This thesis investigated the theoretical processes that underlie perceptual and functional categorisation: perceptual categorisation refers to the process of forming an integrated representation of a pattern of stimulation and functional categorisation refers to the process of integrating otherwise equivalent patterns of stimulation according to their uses or consequences. Investigation of perceptual categorisation in people and of functional categorisation in rats provided results that place important constraints on the nature of the involvement of elemental and configural processes.

In nine experiments rats were required to escape from a swimming pool of opaque water by swimming to a submerged platform. The position of the platform was determined by the shape of the pool, which was either rectangular or triangular. In Chapter 2, a spherical landmark, above the surface of the water, failed to overshadow (Experiment 1) and block (Experiments 2 & 3) learning about the position of the platform relative to these shapes. The same landmark also failed to overshadow learning to find the platform relative to a rectangle described by four identical landmarks (Experiment 4). Experiments in Chapter 3 provided equivalent results with cues situated outside the pool (Experiments 5a & 5b). Experiment 6 revealed that the presence of the triangular-shaped pool potentiated learning based on the cues which surrounded it. In Chapter 4, a landmark was attached to the platform which potentiated learning based on the rectangular pool (Experiment 7). The results of Experiment 8, however, make it unlikely that such potentiation resulted from associations developing between the shape of the pool and the landmark (Experiment 8). These findings offer little support to theories of associative learning like the Rescorla-Wagner model (1972) but are in general agreement with the proposal that animals possess a dedicated Geometric Module, impenetrable to featural information (Cheng, 1986 Gallistel, 1990). The results of Experiment 9 may bring this interpretation of the results into question. In that study, rats were trained to find a submerged platform in one corner of a rectangular, and then a kite-shaped pool. In the kite the platform was easier to find if it was located in a corner congruent with the corner it had previously occupied. Rats must have therefore found the platform with reference to local cues rather than the overall shape of the pool.