New User Registration

In order to register on this site, you must first submit the passphrase below.

2000 - July 20/22 University of Cambridge Talk Abstracts
Article Index
2000 - July 20/22 University of Cambridge Talk Abstracts
Page 2
All Pages

Wednesday July 19, 6pm

The Hebb Award Lecture

Making sense of humours

Jane Stewart

Concordia University

In the study of motivation and emotion, body 'humours' have long played an explanatory role. This talk will be concerned with how our concepts of 'humours' have evolved. Using examples from recent behavioral studies I will discuss how modern knowledge of the effects of the ever increasing numbers of 'humours' and their modes of influence is increasing our understanding of how motivational states are induced, about the role of learning in motivation and about how state variables alter behavior.

Session A - Thursday 20 July 9-10.40 am

Symposium: "Role of knowledge, logic & normative standards in human inference"

Organiser: Valerie Thompson

The role of suppositions in the development of children's reasoning

Simon Handley and Joanna Lott

University of Plymouth

This paper reports the results of two experiments that examined the development of conditional reasoning in six age groups (7,9,11,13,15,17 and adult). Rule-based theories of deduction provide detailed accounts of the developmental stages in conditional reasoning. These theories predict a reduction in Modus Tollens (MT) reasoning in mid to late adolescence as children begin to resist invited inferences but do not possess competence in applying the indirect suppositional strategies necessary to solve MT. However, the evidence for this trend in MT reasoning is equivocal and depends to a great extent on the method of problem presentation. In Experiment 1 we presented conditional syllogisms to six age groups, carefully controlling the presentation of the problems. The results showed a general increase in reasoning competence with age, a reduction in fallacies amongst the older age groups and a significant reduction in MT reasoning at 17 years of age. In Experiment 2 we used the negations paradigm, with conclusions to be evaluated that were either consistent or inconsistent with the logical conclusion. Two notable effects emerged an increase in negative conclusion bias with age and higher rates of MT reasoning amongst the older age groups on those problems in which the inconsistent conclusion was presented. We interpret these findings as resulting from the development of the same underlying competence; the ability to engage in suppositional reasoning. The implications of the findings are discussed in the context of rule-based and model-based theories of human reasoning.

Probabilities and polarity biases in conditional inference

Mike Oaksford1, Nick Chater2 and Joanne Larkin2

1. Cardiff University

2. University of Warwick

A probabilistic computational level model of conditional inference is proposed that can explain polarity biases in conditional inference (e.g., J. St.B. T. Evans, 1993). These biases are observed when J. St.B. T. Evans' (1972) negations paradigm is used in the conditional inference task. The model assumes that negations define higher probability categories than their affirmative counterparts (M. Oaksford & K. Stenning, 1992), e.g., P(not-dog) > P(dog). This identification suggests that polarity biases are really a rational effect of high probability categories. Three experiments revealed that, consistent with this probabilistic account, when high probability categories are used instead of negations a high probability conclusion effect is observed. In the General Discussion the relationships between the probabilistic model and (i) other phenomena and (ii) other theories, in conditional reasoning are discussed.

Paradoxical individual differences in conditional inference

Jo Sellen and Mike Oaksford

Cardiff University

Paradoxical individual differences, where a dysfunctional trait correlates positively with some preconceived notion of the normatively correct answer, provide compelling evidence that the wrong norm has been adopted. We report a result showing that logical performance on conditional inference is positively correlated with schizotypy. Schizotypy is strongly negatively correlated with IQ. Therefore following Stanovich and West's (in press) reasoning, logic may not be normative in conditional inference, the prototypically logical task. This is consistent with recent models of conditional that take a probabilistic rather than a logical approach (Oaksford, Chater & Larkin, 2000).

Knowledge-based reasoning about causes and consequences

Jonathan Fugelsang and Valerie Thompson

University of Saskatchewan

When evaluating causal candidates, peoples' judgments may be influenced by both the observed empirical evidence (e.g., Cheng, 1997), and the belief in a causal mechanism (e.g., White, 1989). Fugelsang and Thompson (2000) presented a series of experiments which demonstrated that people weigh empirical evidence in light of their pre-existing beliefs, such that empirical evidence is given more weight for believable than for unbelievable candidates. The goal of the current experiments was to determine what elements of a reasoner's beliefs produces this interaction. The results of two experiments demonstrated that it is the belief in a causal mechanism, rather than the belief that the cause and effect covary, that determines the use of empirical evidence.

Background beliefs about the probability of the evidence in hypothesis testing

Aidan Feeney1 and Jonathan Evans2

1. University of Durham

2. University of Plymouth

Previous work (Feeney, Evans and Clibbens, 1997; in press) has demonstrated that people's background beliefs about the probability of the evidence affect both their selection and interpretation of evidence on the pseudodiagnosticity task. I will present the results of two experiments designed to investigate the processes underlying such probabilistic effects. Although both experiments replicate the background belief effect in evidence selection, they also demonstrate that hypothesis testers are inflexible in the face of explicit statistical and implicit contextual manipulations of the likely information to be gained from selecting evidence with low a priori probability. It will be argued that these results suggest the operation of a rarity heuristic in hypothesis testing whilst possible adaptive functions for such a heuristic will also be discussed.

Session B - Thursday 20 July 9-10.40 am

Symposium: "It’s about time"

Organiser: Simon Grondin

The time of my life

Lorraine Allan

McMaster University

Allan and Kristofferson (1974) concluded that "There are few quantitative theories of duration discrimination and few established empirical phenomena to guide theorizing." (p. 26). Five years later, Allan (1979), referring back to Allan and Kristofferson (1974), stated that "Since that time, there has been a dramatic change. Many articles on the discrimination of brief temporal intervals, and several new quantitative models, have appeared in a period of a few years." (p. 340). At BBCS/EPS 2000, it will be two decades since Allan (1979). We now have an abundance of sophisticated quantitative models for time perception and a wealth of intriguing empirical data. My contribution to the ITS ABOUT TIME symposium will be a personally-oriented tour through the psychophysical timing literature of the past two decades.

Temporal discrimination and temporal differentiation of behaviour: Evidence for distinct neural mechanisms

C M Bradshaw, M-Y Ho, T-J Chiang, S Mobini and E Szabadi

Psychopharmacology Section, Division of Psychiatry, University of Nottingham

Schedules used to assess timing in animals include 'retrospective timing schedules', in which animals are trained to emit different responses following stimuli of different durations (temporal discrimination), and 'immediate timing schedules', in which behaviour is regulated during an ongoing interval (temporal differentiation). Observations that the phenomena of timing behaviour are similar in these two types of schedule have encouraged the belief that (a) a common timing mechanism underlies temporal discrimination and temporal differentiation, and (b) quantitative indices of timing derived from either type of schedule may be used interchangeably to investigate the properties of this mechanism. However, these assumptions are challenged by evidence that neurobiological interventions can selectively alter timing indices in some schedules but not others. This paper discusses evidence from experiments in which central monoaminergic functions were manipulated by acute drug treatment and lesions; it is proposed that the findings are inconsistent with the concept of a unitary pacemaker-driven internal clock.

It's about time making sense, or perhaps about senses making time

Simon Grondin

Université Laval, Québec

Time judgments are often reported to be based on the output of a central clock. Such a view assumes that for whatever sensory-modality condition under investigation, sensory signals are temporally linked by the product of this central source of time. In this case, time in itself contributes to the making of the sensory organization. On the other hand, psychological time can be approached as an emergent property of the organization between sensory events within any given sensory modality. That is to say, the senses are what make time. Studies stemming from the clock perspective and involving sensory modality comparisons for duration discrimination are briefly reviewed. As well, the implications of adopting a "sense-making-time" approach are presented for one case, that of visual modality.

DMTS as JOR

Peter R Killeen

Arizona State University, USA

Forgetting curves in delayed matching-to-sample (DMTS) experiments are too steep to be credible evidence for rapid forgetting of stimuli that are so very-well known. Instead, animals are deciding whether the remembered stimulus occurred on this trial or on the previous one. This is judgement of recency (JOR). Several models of temporal confusion are reviewed; the simplest, scalar timing, makes parameter-free predictions that approximate DMTS data over various manipulations.

Infernal machines! Exploring internal clocks and other timing processes

J H Wearden

Manchester University

A persistent problem in understanding how the hypothetical internal clock underlying interval timing in humans might work is that the clock is usually supposed to be only the "front-end" of a complex timing system involving memory and decision processes in addition to the clock, where behaviour observed depends on the operation of the whole system. Two complementary approaches to research in this area are (i) to try to "isolate" parts of the system (e.g. the clock) leaving everything else intact, and (ii) to study the way in which the behaviour of the whole system varies across conditions. This paper illustrates both approaches using data from very recent work conducted at Manchester. Firstly, attempts were made to alter clock processes alone by experiments which exploited (a) arousal manipulations and (b) modality effects, including effects of irrelevant stimuli. Secondly, a large-scale study of individual differences in timing behaviour is reported, where nearly 100 subjects each performed on 7 different timing tasks. These were temporal generalization with short and long standards, verbal estimation, and threshold determination [stimulus timing]; production, reproduction, and internal tempo [motor timing]. Correlations between performance indices from the different tasks offer us some insight into the way that the clock/memory/decision structure of the proposed timing system behaves under different conditions.

Session A - Thursday 20 July 11.20-1 pm

Symposium: "Role of knowledge, logic & normative standards in human inference" (continued)

Warnings and inducements as argument

Valerie Thompson1, Jonathan Evans2 and Simon Handley2

1. University of Saskatchewan

2. University of Plymouth

In natural contexts, conditional statements are often used as part of an argument, wherein the conditional premise is used to invite a conclusion. This study examined inferences invited by conditional warnings and inducements. For example, a conditional warning of the form "if tuition fees are raised, then students will no longer be able to afford a university education", appears to invite the inference that "tuition fees should not be raised". Several paradigms were used to examine the suppositions that support the invited inference (i.e., that it is undesirable for students to be unable to afford university education), as well as the chain of logic that supports the invited inference (i.e., if p, then q; want not-q, therefore should not-p). Reasoners were asked a) which inferences followed, b) to generate rebuttals to the invited inference, and c) to evaluate rebuttals that were provided by the experimenter.

Alternatives, falsification, and mental models theory

Steven Newstead1, Simon Handley1 and Valerie Thompson2

1. University of Plymouth

2. University of Saskatchewan

In this experiment, we tested Mental Model Theory's assumption that reasoners construct alternative models of the premises in order to falsify putative conclusions. Participants completed a task designed to measure their ability to construct alternative representations of premises, as well as a syllogistic reasoning task. For the latter task, reasoners were asked to evaluate four types of conclusions: valid conclusions, highly accessible but invalid conclusions (i.e., those that are likely to be considered first as putative conclusions), inaccessible conclusions (i.e., those that are unlikely to be considered as possible conclusions), and impossible conclusions. Reasoners who scored well on the alternative representations task were more likely: a) to correctly accept valid inferences, b) to correctly reject the highly accessible, but invalid conclusions, and c) to correctly identify the inaccessible conclusions as being consistent with the premises. These findings are consistent with the assumption that falsify putative conclusions by searching for alternative representations of the premises.

(End of symposium)

The p-value fallacy: Why inferential statistics don’t describe results

Peter Dixon

University of Alberta

The methods of statistical inference were originally proposed as a mechanical procedure for drawing conclusions from data, but they are rarely used in that way. Instead, inferential statistics are typically used to describe results and how well they match different possible interpretations; scientists draw their own conclusions. This mismatch of methods and goals underlies many of the well-known logical fallacies and interpretive problems in the standard hypothesis testing framework. A better approach is simply to articulate those different possible interpretations and then to describe the match of those interpretations to the data using likelihood ratios. Such an approach provides, in a direct and intuitive form, precisely the information that is conveyed by p values, but without the distracting and problematic rhetoric involved in testing null hypotheses.

Dual task co-ordination versus task difficulty: Evidence from Alzheimer patients and healthy adults

R H Logie1, S Della Sala1, G Cocchini1 and A D Baddeley2

1. University of Aberdeen

2. University of Bristol

The ability of human beings to perform more than one thing at a time has long been a focus of study in the literature on human attention and on memory. However there is a continuing debate as to how the co-ordination of multiple tasks is accomplished, what kind of cognitive functions might be involved, and what determines success or failure in multiple task performance. Experiment 1 demonstrated that AD/pts, healthy elderly and young showed decrements in dual task performance even when each task constituted a very light cognitive load for the individual, but with much larger decrements shown for the patients. However patients were no more sensitive to increased task demand on single tasks than were healthy participants. These results might have arisen from the use of a memory task (span) together with a perceptuo-motor task (tracking). In Experiment 2, healthy adults performed either two memory tasks (visual pattern span and digit span), or a combination of each memory task with tracking. The largest dual task decrement was observed when visual pattern span was combined with tracking. The overall pattern is consistent with a multiple component working memory model, and an identifiable cognitive resource for dual task co-ordination that is insensitive to task difficulty.

A longitudinal study revealing age-related stability in dual-task performance

Margaret Crossley

University of Saskatchewan

Of 92 original participants in a dual-task study of normal aging, 40 were retested in 5-year and 10-year follow-up studies. Of these, 36 remained "healthy" and formed final samples of young (n= 9; ), middle-aged (n = 11), and older (n = 16) participants, with average ages of 40.6 (SD = 5.3), 63.1 (SD = 7.4) and 79.3 (SD = 4.6) years, respectively. At each study period, participants completed neuropsychological tests of memory and attention, and a dual-task study combining speeded finger-tapping with either "easy" (speech repetition) or "difficult" (letter generation verbal fluency) speaking tasks. Although the expected age group differences were evident, tests of memory remained remarkably stable across 10 years for the three age groups, while performance on measures of speeded visual attention declined for all of these healthy participants. Experimental data were analyzed in repeated measures ANOVA, with Age Group as the between- group measure and Study Period (original, 5-year, 10-year) and Task Difficulty ("easy" versus "difficult" speech tasks) as within group repeated measures. Single-task performance of speeded finger-tapping, speech repetition, and speech fluency declined across the 10-year study period for all age groups, and significant Study Period X Age Group interactions reflected greater loss across time for the older groups. As expected, interference during dual-task tapping (controlling for single-task performance) was greater for the oldest group at each test period, and was greater while concurrently performing the "difficult" speaking task compared to the "easy" speaking task. In contrast to expectations, interference did not increase across Study Period, and Age Group did not interact with Study Period or with Task Difficulty. These results suggest that speed of processing, but not attentional capacity, declines with age, at least as measured by these dual task data collected from healthy participants over a 10 year study period.

Session B - - Thursday 20 July 11.20-1 pm

Stimulus modality and the perception of empty time intervals in pigeons

Angelo Santi, Lori Ross and Andrew Miki

Wilfrid Laurier University, Ontario

Pigeons were trained to discriminate the duration (2 or 8 s) of an empty interval separated by two tone markers or by two light markers. Psychophysical testing indicated less sensitivity to empty intervals signalled by tone than by light. Pigeons judged the intervals marked by light to be longer than the same interval marked by tone. Training and psychophysical testing was carried out with empty intervals at target durations of 1 and 4 s, or 4 and 16 s. Pigeons continued to judge intervals marked by light to be longer than the same intervals marked by tone. The difference in the point of subjective equality for light and tone marked intervals was not constant across the different target duration sets. This suggests that the modality difference in the timing of empty intervals is not due to a difference in the latency to initiate timing.

Symposium: "Development of motion perception"

Organiser: Terri L Lewis

Processing of motion information in the primate cerebral cortex.

J Anthony Movshon

Howard Hughes Medical Institute at New York University, USA

A small group of areas in the primate cerebral cortex seems to be specialized for the processing of information about visual motion. Starting with signals from motion-sensitive cells in primary visual cortex, these areas contain a concentration of cells sensitive to motion. Motion-sensitive cells in V1 are primarily concerned with the representation of local first-order motion; this information is progressively elaborated by neurons in areas MT and MST into representations of global motion and optic flow. The outputs of MT and MST provide signals to areas responsible for perceptual judgements of motion and for the initiation of behavior dependent on motion information, such as pursuit eye movement.

While such key basic properties as directional selectivity appear very early in postnatal development and can be disrupted by some forms of visual deprivation, the maturation and sensitivity to environmental manipulation of more complex stages of motion analysis are not well understood at the neurophysiological level.

Interacting motion processing mechanisms in the developing infant.

Oliver Braddick, John Wattam-Bell, Alexandra Mason and Janette Atkinson

Visual Development Unit, University College London

We find no evidence from preferential looking (PL) of directionally selective processing before 7 weeks of age. However, directional mechanisms must underlie the newborn's optokinetic nystagmus (OKN) response. Parallel measurements of motion coherence thresholds, using OKN and PL, show that the two responses differ greatly in developmental course and sensitivity between 8-25 weeks of age.

These results are consistent with a model in which PL reveals the development of cortical motion processing, while neonatal OKN depends on a subcortical circuit, whose signature is temporal/nasal asymmetry in monocular OKN. Developmental changes in children with extensive unilateral lesions show that this initial subcortical asymmetry is modified by cortical-subcortical interactions.

Asymmetries also appear in evoked-potential (VEP) and PL responses. We find that the PL asymmetry follows the OKN asymmetry, while the VEP asymmetry is opposite. These relationships will only be understood in terms of developmental interactions between cortical motion processing and oculomotor mechanisms.

What does blindsight tell us about the functional architecture of motion perception?

Paul Azzopardi1, Mazyar Fallah2, Charles Gross2, Hillary Rodman3 and Alan Cowey1

1. University of Oxford

2. Princeton University, USA

3. Psychology Department and Yerkes RPRC, Emory University, USA

Patients rendered clinically blind as a result of striate cortex damage may nevertheless be able to discriminate stimuli presented in their field defects when forced-choice procedures are used. But there are conflicting reports as to whether or not the ability to discriminate motion is preserved. We recorded the responses to moving stimuli of neurons in areas MT and MST of monkeys with long-standing unilateral striate cortex lesions. Many neurons responded to the presence of movement in the scotoma, and some could discriminate the direction of single bars, but none could discriminate the direction of motion in random dot kinetograms depicting translation or motion-in-depth. Subsequently, we tested the ability of cortically-blind patients to discriminate the same stimuli, with the same results. Direction discrimination in the scotoma seems to be based solely on change in position, implying that motion processing per se is abolished. Blindsight therefore provides no evidence to support the proposed existence of a subcortical pathway to extrastriate cortex which bypasses striate cortex and which is specialised for analysing "fast" motion.

The perception of local and global motion after early pattern deprivation in humans.

Terri L Lewis1, 2, 3, Dave Ellemberg2, Daphne Maurer1,2, Nancy Defina2 and Henry P Brent1,3

1. The Hospital for Sick Children, Toronto

2. McMaster University

3. University of Toronto

We assessed the perception of local and global motion in patients who had been deprived of early visual experience by dense central cataracts. Despite years of visual experience after treatment, patients showed deficits on both tasks. Like our previous findings for contrast sensitivity, the deficits in sensitivity to local motion were comparable after either monocular deprivation (n = 10) or binocular deprivation (n = 9). Surprisingly, for global motion tested with random-dot kinematograms, coherence thresholds were nearly normal (M = 8.5%, n = 24) after monocular deprivation (M = 14%, n = 14) and significantly worse after binocular deprivation (M = 44%, n = 8). The unexpectedly good results after monocular deprivation suggest that normal input to only one eye reduces the deleterious effects of deprivation on the higher cortical areas involved in the perception of global motion. Thus, after monocular deprivation, competitive interactions occurring in the primary visual cortex may give way to collaborative interactions at higher neural centres enabling a relative sparing of at least some visual functions.

(End of symposium)

Session A - Thursday 20 July 2-4.20 pm

Temporal jitter disrupts speech intelligibility: Simulations of auditory aging

M. Kathleen Pichora-Fuller1, Bruce Schneider2 and Hollis Pass1

1. University of British Columbia

2. University of Toronto

Auditory temporal processing declines in aging adults may partially account for their difficulty understanding speech. To mimic such age-related declines, sentences were temporally jittered. Intact and jittered sentences were presented in babble to young listeners. For the signal x(t), the internal representation y(t) is modeled as a time-delayed version with delay d varying over time as y(t)=x[t-d(t)]. The first factor is the range of delays, modeled as the RMS of a band-limited noise. The second factor is rate of change of delays, modeled as the bandpass of a noise. Word recognition was unchanged when the second factor dominated (RMS = .05 ms, BW = 500 Hz), reduced slightly (10%) when the first dominated (RMS = .25 ms, BW = 100 Hz), and reduced greatly (50%) when both contributed (RMS = .25 ms, BW = 500 Hz). This is a first attempt to simulate temporal aspects of auditory aging.

Symposium: "Cognitive aging"

Organisers: Patrick Rabbitt and Fergus Craik

Age slows, and so sometimes obliterates consciousness of events.

Patrick Rabbitt

University of Toronto

After people make errors during simple continuous tasks they recognise them at some level because their sequent responses are slowed. People can correct nearly all of the errors that they make during simple continuous tasks by making the response that they should have made.Old age does not affect the efficiency of either of these forms of error detection and correction. However when people are asked to signal their errors by immediately making some other response they do so slowly and inaccurately, and the older they are the fewer responses they signal. When probed, they do not recall making errors. Nevertheless, even those errors that are not signalled are invariably followed by slow responses to sequent events. It seems that events that are not overtly acknowledged are still registered, but not remembered. Increasing the intervals between successive signals and responses improves performance, so that when these are as long as 2 sec error signalling is as efficient as error correction. The age difference in efficiency of error signalling disappears. We suggest that this means that simple events have successive, perhaps independent , representations which become serially available in time. During a fast paced task atention to new signals and responses prevents attention to later arriving representations. Because all cognitive processes are slowed by age older people need more time to become conscious that they have made an an error and to do something about it. Self-monitoring, even in simple tasks, is slowed by age, and thus becomes more vulnerable to the pace of events.

Aging and visual marking: Selective deficits for moving stimuli.

Elizabeth A Maylor and Derrick G Watson

University of Warwick

Previous research has shown that new visual information can be selectively prioritized by the top-down inhibition of old stimuli already in the visual field (visual marking; Watson & Humphreys, 1997). Moreover, the method of visual marking depends upon the properties of the old items such as their motion and the extent to which they can be grouped. We present three experiments to assess the effects of aging on visual marking. Experiment 1 examined performance with stationary items; Experiments 2 and 3 used moving stimuli. The results showed that visual marking was relatively unimpaired by aging for stationary stimuli. In contrast, there were age-related deficits in visual marking for moving stimuli, particularly when there was no feature difference between the old and new items. These differential effects of aging provide support for the deployment of different methods of visual marking depending upon the properties of the old (to-be-ignored) items.

Watson, D. G., & Humphreys, G. W. (1997). Visual marking: Prioritizing selection for new objects by top-down attentional inhibition of old objects. Psychological Review, 104, 90-122.

Adult-age differences in memory performance: Tests of an associative deficit hypothesis

Moshe Naveh-Benjamin

Ben-Gurion University of the Negev, Israel

An associative hypothesis to explain and predict older adults' deficient explicit episodic memory performance is outlined and tested. The hypothesis attributes a substantial part of older adults' deficient memory performance to their difficulty in merging unrelated attributes-units of an episode into a cohesive unit. While each of the components can be memorized to a reasonable degree, the associations which tie the attributes-units to each other grow weaker in old age. Four experiments are reported which provide: a) a converging validity to the hypothesis by demonstrating this associative deficit for both inter-item relationships and intra-item relationships, and b) a discriminant validity to the hypothesis, by contrasting and testing competing predictions made by the associative hypothesis and by alternative hypotheses. The implications of these results to older adults’ episodic memory performance are discussed.

Retrieval induced forgetting in normal aging.

Tim Perfect1, Chris Moulin2 and Martin Conway3

1. University of Plymouth

2. University of Reading

3. University of Bristol

In the field of cognitive aging it has been suggested that the cognitive deficits experienced by older adults stem from a general deficit in inhibitory function. In memory research, the retrieval-induced-forgetting paradigm has been used to isolate inhibitory processes that occur during retrieval. In a series of studies we examine whether there are age-related reductions in inhibition during memory retrieval. Younger and older adults were tested on a series of explicit and implicit tasks using the retrieval induced forgetting paradigm. None of the studies provide evidence in support of our predictions. These data suggest either that inhibitory effects are universal across the age range, or inhibitory mechanisms are not involved in retrieval. These two possibilities will be discussed.

Adult aging and planning ability

Louise Phillips and Mairi MacLeod

Aberdeen University

The formulation and execution of efficient plans is dependent upon the prefrontal cortex of the brain. Normal adult aging causes changes in the structure and function of the frontal lobes, but relatively little is known of the nature of age effects on planning. The current experiment looked at the effects of aging on a multistage task of party organisation in 96 participants aged between 16 and 79. Older participants showed poorer planning, and tended to make more rule breaks such as trying to buy items with insufficient funds. There were no age effects on other types of error, such as interpretation failuresor inefficiencies of task allocation. A qualitatively different pattern of errors was seen in young and older groups. These results can be contrasted with the pattern of age differences on a series of other tasks of planning (the Tower of London task, Multiple Errands test, and subtests from the Behavioural Assessment of Dysexecutive Syndrome). Overall, the results suggest that the nature and extent of age differences in planning vary considerably depending on such factors as: memory load, perceived importance of rules in the task, and planning complexity

(End of symposium)

Session B - Thursday 20 July 2-4.20 pm

Auditory attentional blink in the presence or absence of filler items

Sébastien Tremblay and Dylan M Jones

Cardiff University

Identification of the second of two targets in a rapidly presented sequence of stimuli is often incorrect when it falls within about 500 ms of the first target. Although the phenomenon of attentional blink (AB) is well-established in the visual modality, some claim that there is no AB for auditory sequences (Potter, Chun, Banks & Muckenhoupt, 1998) while others report a robust auditory AB (Arnell & Jolicoeur, 1999; Duncan, Ward & Shapiro, 1994). The present study provides more evidence that there is an effect of AB in the auditory modality. There was a significant deficit in target two report accuracy but it diminished as a function of the temporal lag between the two targets. The magnitude of the auditory AB was greater when there were filler items in between the targets but even in the absence of fillers the AB was significant. The auditory AB reported here was not due to task switching deficit as the same discrimination task applied for both targets.

Attention, occlusion, and previous experience

Jay Pratt and Allison B Sekuler

University of Toronoto

There is considerable evidence indicating that cueing a specific portion of an object results in the entire object being attended to. One goal of the present study was to determine whether attentional cueing effects act on objects defined only in terms of occlusion cues. The second goal was to determine whether previous experience with an object could halt perceptual (i.e., amodal) completion. In Experiment 1, two parallel rectangles were initially displayed and then the middle portions of these objects were occluded. Attentional cueing effects were found for both discrete portions of the completed rectangles. In Experiment 2, four discrete objects were initially displayed, followed by the same occluder used in the first experiment. The appearance of the occluder allowed the four discrete objects to be completed into two rectangles. Attentional cueing effects were found for the completed rectangles, indicating that previous experience may not be sufficient to halt the amodal completion of objects.

Spatial modulation of intramodal and crossmodal temporal processing deficits

S Soto-Faraco1,2, C Spence2, A Kingstone1 and J Duncan3

1. University of British Columbia

2. University of Oxford

3. MRC Cognition and Brain Sciences Unit, Cambridge

The attentional blink (AB) and repetition blindness (RB) effects both reflect costs in the processing of the second of two targets presented in a rapid sequence. The AB is a decrease in detection or identification accuracy on any target following the selection of a previous stimulus in a time window of 200 ms and 500 ms. The time course of AB often takes the form of a u-shaped function (produced by the so called lag-1 sparing). The RB reflects the cost in processing the second of two identical targets and its time course is linear (accuracy is inversely proportional to the temporal separation between targets). Although these two temporal processing deficits are similar in many ways, Chun (1997) demonstrated that they could be dissociated in terms of their differential sensitivity to target-distractor and inter-target similarity. We report a series of experiments in which we attempted to dissociate these effects in terms of their sensitivity to spatial uncertainty, and spatial displacement of the target stimuli. Pairs of targets (either the same or different) were presented from the same or different spatial locations in the same or different modalities. We found both visual and auditory AB and RB effects. Additionally we show that temporal processing deficits are modulated by spatial displacement between target stimuli.

Is perception, attention or action inhibited following spatial reorienting?

Raymond Klein, Bruce Dick, Jason Ivanoff and Tracy Taylor

Dalhousie University, Halifax

Immediately following a peripheral event there is a facilitation for the processing of nearby events. When shifts of gaze are not made this is said to reflect exogenous orienting of attention toward the source of stimulation. After attention is removed from such a source, there is impaired processing of subsequent, nearby events . The name originally given to this processing impairment, inhibition of return (IOR), implies that it is caused by removal of attention from the cued location and that its effect is to discourage attention from reorienting to the originally attended region of space. Despite this implication, whether perceptual, attentional or motor stages of processing are inhibited has been much debated. Recent evidence from our laboratory for effects of IOR on perception, attention and action will be used to illustrate the manifold nature of IOR.

A facilitatory effect does not contribute to the inhibition of return effect

Janice J Snyder1, William C Schmidt2 and Alan Kingstone3

1. University of Alberta

2. University at Buffalo, USA

3. University of British Columbia

It is well established that the "inhibition of return" (IOR) effect is found at a previously attended/cued location at long cue-target onset asynchronies. More recently, it has been reported (Pratt, Spalek, & Bradshaw, 1999) that in addition to the inhibitory effect at the cued location a co-occurring facilitatory effect occurs at the uncued location opposite to the attended/cued location. This finding led to the suggestion that facilitation at the uncued/opposite location may represent the primary component underlying IOR. In three experiments, we attempted to expand on these findings by defining the conditions necessary for observing the opposite facilitation effect. Our data revealed that unlike IOR, the opposite facilitation effect is fragile and weak, demonstrated by only a minority of subjects, and is statistically uncorrelated with the IOR effect. Together the data suggest that an opposite facilitation effect does not contribute significantly to the IOR effect.

Is task-set inhibition in set switching location specific?

Katherine Arbuthnott1 and Todd Woodward2

1. University of Regina

2. University of British Columbia

Set-switching tasks are used to study executive control processes involved in shifting attention between different goals. Costs attributable to task-switching are greater for alternating switches (ABA) than for non-alternating switches (CBA), an effect that is observed for perceptual and semantic judgment tasks, with manual or verbal responses (Arbuthnott & Frank, 2000; Mayr & Keele, 2000). In these investigations, stimuli were presented at a consistent location and the relevant task for each trial was cued verbally. The present studies examined switch costs for alternating and non-alternating series when the tasks were cued by spatial location. When tasks were uniquely associated with widely separated locations, switch cost was greater than with verbal cues but additional task-alternation cost was not observed. This suggests that different processes may be involved when switching among goals that have distinct locations versus goals that must be completed at a single location.

Is egocentric space automatically encoded?

Sandra Pouliot and Sylvain Gagnon

University of Quebec

Parkin, Walter and Hunkin (1995) suggested that egocentric space is automatically encoded. Using the criteria defined by Hasher and Zacks (1979), Naveh-Benjamin (1988) demonstrated that spatial information encoding was not an automatic process. However his spatial memory task probably emphasized encoding of allocentric space. In the present study, we tested Parkin et al's hypothesis using a task that was specifically designed to assess encoding of egocentric space. Five studies analyzing the effects of intent of memorization, dual task interference, old age, practice and individual differences were carried out using a computerized egocentric spatial memory task. Results show that old age and dual task interference impair egocentric spatial memory. However intent of memorization, practice and individual differences did not influence performance. The findings demonstrate that encoding of egocentric space is not a pure automatic process and therefore requires some cognitive effort.

Session A – Friday 21 July 9-10.20 am

Symposium: "Neuroimaging of memory"

Organiser: Roberto Cabeza

The role of the left prefrontal cortex in encoding of episodic memory

Paul Fletcher

Research Department of Psychiatry, University of Cambridge

A growing number of functional neuroimaging studies provide evidence for the importance of left prefrontal cortex (PFC) in episodic memory. The specific role of this region, however, remains unclear. Nevertheless, a number of characteristic features have emerged to suggest that the left PFC shows greater levels of activation when the encoding results in subsequent successful retrieval and when encoding occurs in association with a task which emphasises the semantic processing of studied material. With regard to the involvement in semantic processing demands, three related views have been described. It has been suggested that the left PFC activation reflects (i) the generation of semantic attributes of study material, (ii) the holding of those semantic attributes in working memory and (iii) the selection from among different semantic attributes. Distinguishing between these possibilities is difficult but it is suggested that a number of recent studies favour the latter interpretation. These studies will be reviewed.

Involvement of prefrontal regions on episodic memory retrieval: Mode, success, effort, and GRAM

Roberto Cabeza

University of Alberta

One approach to investigate the role of the prefrontal cortex (PFC) in episodic memory retrieval is to determine whether PFC activity remains constant (retrieval mode), increases (retrieval success), or decreases (retrieval effort) as a function of recovery. In an event-related fMRI study, we associated bilateral dorsolateral PFC with retrieval success, and left ventrolateral PFC with retrieval effort. Left PFC activity during episodic retrieval could reflect semantic processing. Since semantic processing is critical for the generation of candidate responses, left PFC activity should be more pronounced for recall, which requires generation, than for recognition, which does not. A PET study confirmed this prediction, and additionally revealed greater right PFC activity for recognition than for recall. To described this pattern we propose a generate/recognize asymmetry model (GRAM): left PFC is differentially more involved in generation processes than is right PFC, whereas right PFC is differentially more involved in recognition processes than is left PFC.

The functional organisation of working memory processes within the lateral frontal cortex.

Adrian M Owen

MRC Cognition and Brain Sciences Unit, Cambridge

Evidence is now converging to suggests that, at the area level, working memory processes within the dorsolateral and ventrolateral frontal cortices are organized according to the type of processing required, rather than according to the nature (i.e. modality), of the information being processed, as has been widely assumed. In a recent study using fMRI, performance of visual spatial and visual non-spatial working memory tasks was shown to involve identical regions of the lateral prefrontal cortex when all factors unrelated to the type of stimulus material were appropriately controlled. Moreover, recent positron emission tomography studies have demonstrated that either, or both, the ventrolateral and dorsolateral frontal regions can be activated in spatial, visual or verbal working memory tasks, depending on the precise (non-modality specific) executive processes that are called upon by the task being performed. The results provide further evidence that the mid-dorsolateral and mid-ventrolateral frontal cortical areas make distinct functional contributions to memory and corresponds with a fractionation of working memory processes in psychological terms.

Age-related differences in brain activity during memory

Cheryl L Grady

Rotman Research Institute, University of Toronto

Neuroimaging experiments have shown that brain activity patterns in older adults during performance of cognitive tasks differ from those seen in younger adults, even if their ability to do the task is not markedly different. Older adults have activation in many of the same areas as do young adults, but also have greater activity in other brain regions, particularly in prefrontal cortex. Various interpretations could be made of this finding, including compensatory mechanisms, or lack of inhibitory processes. In order to distinguish among these possibilities we need to determine how these age-related changes in brain activity are related to behavior in old adults, e.g. whether increased activity in a specific brain region occurs when task performance is spared in an older group or whether increased activity is associated with better or worse performance on a specific task. Recent experiments that shed light on this issue will be reviewed.

(End of symposium)

Session B - Friday 21 July 9-10.20 am

The resource demands of semantic priming

Eric Richards and Jennifer Stolz

University of Waterloo

The resource demands of semantic priming were examined in a series of experiments using a modified dual task paradigm. A prime word preceded a target stimulus, which was temporally paired with a tone. Participants were required to categorize the pitch of the tone prior to making a decision about the target stimulus (i.e., semantic classification or lexical decision). Using the logic that the resources required to process the target would limit those available for tone processing (Becker, 1976; Herdman & Dobbs, 1989), tone RTs were compared in related versus unrelated priming conditions to examine whether resource benefits are associated with semantic priming. The results are discussed in terms of the general theories of word recognition and the capacity constraints of attentional processing.

Binding action to a source: Evidence from Stroop priming

Bruce Milliken1, Juan Lupianez2, Karmen Bleile1 and Jason Leboe1

1. McMaster University

2. Universidad de Granada, Spain

Stroop interference is widely presumed to occur because of the predominance of word reading over colour naming processes. Consequently, priming of the color dimension might reasonably be predicted to reduce Stroop interference. In contrast to this prediction, we report the results of several experiments which demonstrate that priming of the color dimension speeds Stroop performance under some conditions, but slows Stroop performance under others. Attentional factors implicated in this processing flexibility are the focus of the experiments reported. The results are consistent with the view that priming introduces a source of influence on performance that can be difficult to discriminate from the normally fluent processing of the word dimension of Stroop stimuli.

Searching for threat

Jason Tipples1, Andy Young1, Philip Quinlan1, Paul Broks2 and Andy Ellis1

1. University Of York

2. University Of Sheffield

In a series of experiments, a modified version of the visual search task was used to test the idea that biologically-relevant threat stimuli might capture visuo-spatial attention. In Experiment 1, no overall reaction time or search rate advantage was found for angry faces compared to happy faces. In Experiment 2, however, there was evidence for both faster detection and search rates for threatening animals, compared to plants. Examination of the basis of this effect in Experiment 3, showed that it was not due to threat per se, since a detection and search rate advantage was also found for pleasant animals compared to plants. In Experiment 4, participants searched for the plants, pleasant and threatening animals used in Experiments 2 and 3, among a fixed heterogeneous selection of distractor items. There was no search rate or detection advantage for threatening animals compared to pleasant animals or plants. The same targets and distractors used in Experiment 4 were also used in Experiment 5. However, in Experiment 5, participants kept their eyes fixed on a central cross while searching for targets either close or distant from the fixation cross. There was no evidence for a ^Ìthreat¹ detection advantage either close or distant from the cross. We conclude that the visual search paradigm does not readily reveal any attentional biases that might exist for threatening stimuli.

Modulation of specific processes in task switching

Michael E J Masson, Daniel N Bub, and Todd S Woodward

University of Victoria

Task switching is typically assumed to involve executive control in coordinating the change in procedures as a function of change in task demands. A number of studies have examined general switch costs incurred in the performance of the second task as a consequence of having just carried out the first task. In contrast, we describe a process-specific influence of task switching on the performance of the first task. Considerable slowing is observed in the first task when that task shares specific operations with the second task. We have analyzed this phenomenon in the following situation: Task 1 requires reading aloud a word printed in black; Task 2 requires naming the color of a subsequent stimulus immediately following Task 1. On one block of trials, the stimulus in Task 2 is a row of colored asterisks and in another block it is a colored word. We find that word reading in Task 1 is substantially slowed when Task 2 involves colored words as opposed to colored asterisks. This effect does not obtain, however, when Task 1 requires a lexical decision rather than reading aloud, indicating that this modulatory effect is process specific.

Session C - Friday 21 July 9-10.20 am

Assessing inhibitory control in children and adults with a picture choice task

Sandra A Wiebe and James M Clark

Institute of Child Development, University of Minnesota, USA

Inhibitory control develops relatively late in childhood relative to other cognitive functions. Analysis of the development of inhibition would be facilitated by examining change in performance on one task across a wide age range. We have developed a task that is suitable for use with participants ranging from preschool-age to adulthood. Participants are presented with pairs of pictures and are asked to choose the picture that is related to a cue word on one of two dimensions (semantic or phonological). On half of the trials the incorrect picture choice is related to the cue word on the other dimension, and thus the choice of this response must be inhibited. Children show evidence of poor inhibitory control, making fewer correct responses on trials requiring inhibition. Adults successfully make the correct choice, but reaction times are lengthened when inhibition is required. These findings also have implications for our understanding of picture processing.

Intervention in early word learning: An experimental approach

Graham Schafer (Introduced by Professor D C Berry)

University of Reading

A lively debate in language development concerns the existence of stages in lexical learning. It is hypothesised that, before the attainment of certain 'principles', word learning is slow and effortful. As a corollary, infants' first words will be those most afforded by their immediate environment, and modification of the early lexicon (by intervention) will be difficult.

This study investigated modification of the early lexicon by parental intervention. Parents of nine-month-olds were given photographs of objects and animals, and were instructed to use these with their children, for 10 minutes per day, three times per week, in simple tasks (naming, matching, and sorting). After 12 weeks, infants' receptive comprehension was tested in a preferential looking task. Infants were instructed to look at images, which depicted novel exemplars of words from the training set. Infants in the experimental group looked at the images as instructed, while infants in a non-intervention control group did not. It is concluded that small amounts of training can modify early vocabulary. The theoretical implications of this are discussed.

Mismatch negativity and auditory backward recognition masking performance in people with a specific language impairment.

G M McArthur and D V M Bishop

University of Oxford

An influential hypothesis attributes the verbal problems of children with a specific language impairment (SLI) to an impairment in processing rapid or brief sounds. However, some studies have failed to demonstrate such a deficit. The rapid auditory processing skills of 16 people with an SLI and 16 people with normal spoken language skills were assessed with a modified version of Winkler, Reinikainen and Naatanen's (1993) paradigm to determine whether the discrepancies in the literature reflect differences in the sample studied, or methodological factors. This paradigm integrated mismatch negativity (MMN) and auditory backward recognition masking (ABRM) in four conditions that presented trials composed of a standard (25-ms 600-Hz) or deviant (25-ms 700-Hz) test tone followed by a 55-ms 1000-Hz backward masking tone after an interstimulus interval (ISI) of 20, 50, 150, or 300 ms. A fifth condition assessed frequency discrimination by presenting the standard and deviant test tones alone (i.e., without a backward masking tone). We will present data relevant to the rapid auditory processing deficit hypothesis, which predicts that the MMN component and ABRM performance of people with an SLI will be impaired in the short ISI conditions (i.e., 20 and 50 ms), but not the long ISI (150 and 300 ms) or frequency discrimination conditions.

"Yesterday I stroke a horse": Can children’s problems with verb morphology be explained in terms of a low-level auditory deficit?

Dorothy Bishop, Courtenay Frazier Norbury and Josie Briscoe

University of Oxford

Children with specific language impairment (SLI) have disproportionate difficulty with certain aspects of verb morphology. For instance, they frequently omit past tense –ed endings, though when they do produce such inflections they are used appropriately. Debate has centred on whether poor mastery of inflections is a secondary consequence of low-level auditory perceptual impairment, whether it reflects specific impairment in a specialised syntactic module, or whether the computational demands of inflectional morphology overload a limited processing capacity. The first of these possibilities is hard to accommodate with data we shall present comparing the performance of 19 children (aged 5; 9 – 10; 7) with mild–moderate sensorineural hearing impairment (SNH), 20 children with SLI (aged 7; 2 – 13; 0), and age-matched and language-matched groups of control children. On average the SNH group outperformed the SLI group and were comparable to controls. However, a minority of the SNH children were impaired on these tasks. These children were significantly younger than the rest of the SNH group, suggesting that acquisition of finite verb morphology may be delayed in children with hearing impairments, but they appear to catch up over time and do not show the persisting difficulties seen in SLI. Children with SLI demonstrated some ability to use past tense and 3rd person singular inflections correctly, and properties of verbs such as verb frequency and phonological complexity affected their ability to produce appropriate inflections, suggesting that their problems were not solely explicable in terms of syntactic factors, but may be influenced by other processing demands.

Session A - Friday 21 July 11-1 pm

Symposium: "Temporal lobe function in episodic and semantic memory"

Organisers: Karalyn Patterson and Kim Graham

Perceptual-mnemonic functions of the perirhinal cortex in macaques

Elisabeth A Murray

National Institute of Mental Health, Bethesda, MD, USA

The perirhinal cortex, a strip of cortex located in the ventromedial temporal lobe in macaques, plays a critical role in certain types of memory. The talk focuses on three aspects of the mnemonic contributions of the perirhinal cortex. First, the perirhinal cortex has been found to be essential for object recognition and for relating together the different sensory features of objects, thereby facilitating object identification. It carries out these functions even in the absence of the amygdala/hippocampus. Second, the perirhinal cortex serves as the final stage in a ventral visual cortical processing stream, known as the "what" pathway, that is devoted to the perception and identification of environmental stimuli. Its special contribution to this type of processing is held to be in the representation of complex conjunctions of features. Thus, the perirhinal cortex participates in acquisition, retrieval, and long-term storage of knowledge about objects, and, together with other cortical fields, also serves as a site of long-term storage of such knowledge. Third, the perirhinal cortex conveys already-processed information about objects to the hippocampus and amygdala, thereby allowing these structures to make different, secondary contributions to object memory, associating objects with events and affective valences, respectively.

Remote memory and the hippocampal complex in humans

Morris Moscovitch

University of Toronto at Mississauga

The studies address the question of whether the role of the hippocampal complex in memory is temporally-limited, needed only until memories are consolidated elsewhere, or whether it continues to contribute to storage and retrieval even of remote memories. Evidence will be presented from studies of people with hippocampal complex lesions, from a person with semantic dementia, and from neuroimaging studies in normal people that indicates that the hippocampal complex is needed for recollection of details of remote autoiographical episodes, public events, and even spatial memories, but not of semantic memories. The results call for a modification of traditional consolidation theory. A multiple-trace theory of memory ( Nadel and Moscovitch, 1997) provides a viable alternative.

Semantic dementia: A challenge to the multiple-trace model of memory consolidation?

Kim S Graham

MRC Cognition and Brain Sciences Unit, Cambridge

For many years, it has been thought that the hippocampal complex is only involved in the retrieval of recently experienced episodic memories. Evidence to support this view comes from patients with bilateral hippocampal damage who show better recall of distant memories compared to more recent personal events. An alternative theory proposes, however, that the hippocampus is important for the retrieval of all personal event memories, regardless of their age (Nadel and Moscovitch, 1997). Here, these two models of memory consolidation will be evaluated by reviewing neuropsychological data from semantic dementia: these patients, who have focal atrophy of the temporal neocortex with relative sparing of the hippocampal complex, have a relatively selective, yet progressive, loss of semantic memory. In support of a temporary role of the hippocampus in episodic recall, patients with semantic dementia show better retrieval of recent autobiographical memories compared to those from the more distant past. To account for these contradictory results, Moscovitch and Nadel (1999) have proposed a number of possible explanations: (1) too few replications of the data; (2) impaired strategic retrieval due to concomitant frontal damage; and (3) preserved non-verbal access to autobiographical memory. Further data from semantic dementia will be considered in order to address the possibility that one or other of these factors explains the observed advantage for recent personal events.

 

Moscovitch, M. and Nadel, L. (1999). Multiple-trace theory and semantic dementia: Response to K.S. Graham (1999). Trends in Cognitive Sciences. 3, 85-122.

Nadel, L. and Moscovitch, M. (1997). Memory consolidation, retrograde amnesia and the hippocampal complex. Current Opinion in Neurobiology. 7, 217-227.

Dissociations in cognitive memory: The syndrome of developmental amnesia.

Faraneh Vargha-Khadem1, David G Gadian1 and Mortimer Mishkin2

1. Institute of Child Health, University College London

2. National Institute of Mental Health, Bethesda, USA

The dearth of studies on amnesia in children has led to the assumption that when damage to the medial temporal lobe system occurs early in life, the compensatory capacity of the immature brain rescues memory functions. An alternative view is that such damage so interferes with the development of learning and memory that it results not in selective cognitive impairments but in general mental retardation. Data will be presented to counter both of these arguments. Results obtained from a series of 12 amnesic patients with a history of hypoxic ischaemic damage sustained perinatally or during childhood indicate that regardless of age at onset of hippocampal pathology, there is a pronounced dissociation between episodic memory, which is severely impaired, and semantic memory, which is relatively preserved. A second dissociation in these patients is characterized by impaired recall and preserved recognition. These findings are discussed in terms of the neuropathology associated with hypoxic ischaemic damage and a hierarchical model of cognitive memory.

(End of symposium)

Session B - Friday 21 July 11-1 pm

Crossmodal links in covert endogenous spatial attention between audition and touch

Donna M Lloyd

University of Manchester

Three experiments were designed to investigate crossmodal links in endogenous spatial attention between audition and touch, using the orthogonal spatial cuing paradigm. Participants discriminated the elevation (up vs. down) of auditory or tactile targets presented to either the left or right. Experiment 1, in which targets were expected on a particular side in just one modality, demonstrated that people could shift their attention around independently in either audition or touch. Experiment 2 demonstrated that when people were informed that targets were more likely on one side for both modalities, elevation judgments were faster on that side in both audition and touch. However, it was possible to ‘split’ auditory and tactile attention, when targets in the two modalities were expected on constant but opposite sides throughout a block. A final study with crossed hands suggests that audiotactile links in spatial attention apply to common external locations, rather than simply being determined by the hemisphere to which information initially projects. These results are discussed in relation to previous findings regarding audiovisual and visuotactile links in attention, and to their possible neural substrates.

Symposium: "Confidence in human judgement"

Organisers: Joseph V Baranski and Nigel Harvey

Moderation of base rate neglect by group discussion and range specification: A calibration study

Tim Rakow, Sarah Finer and Nigel Harvey

University College London

Medical students judged the likelihood that an applicant to medical school would receive an offer of a place on the basis of 8 cues. Participants were naïve to the task, though familiar with its context. In experiment 1 participants made individual judgements before and after group discussion. Consistent with the literature on lay judgement, participants were poorly calibrated and far less accurate than a logistic regression model. Poor calibration was largely attributable to considerable 'bias', with 'discrimination' being comparable to the statistical model. The simple provision of base rate information (the proportion of applicants receiving offers) had minimal impact upon the magnitude of bias of individual judgements. Bias was significantly reduced by group discussion when base rate information was provided (with no detriment to discrimination), but remained unaltered in the absence of base rate information. In experiment 2 some participants were provided either information concerning the distribution of probability estimates or the range of estimates generated by the statistical model. The participants receiving this information were well calibrated, and their accuracy was close to that of the logistic regression model. Results suggest our participants could identify and use appropriate cues, but needed assistance in the scaling of their response. Specifying a 'plausible range' for responses was sufficient to foster well-calibrated judgement. Extensive feedback or instruction in cue utilisation was not necessary in this instance.

The hard-easy effect in the calibration of subjective probabilities: The effect of

defining 'difficulty' in terms of absolute difference rather than familiarity

Alastair McClelland, John Haynes and Zhuo Jia Sun

University College London

Juslin (1993) obtained good calibration at different levels of performance when items were segregated on the basis of rated familiarity, rather than solution probability. In the present study, we report a partial replication and extension of Juslin's work. In the first experiment, participants were presented with a list of country names, and rated each country for familiarity. They were then given pairs of country names (selected at random) and had to decide which country had the bigger land area. For each pair, a half-scale confidence rating was also requested. The items in each domain were then split into two categories on the basis of either average familiarity (following Juslin) or absolute difference in land area. With familiarity, the hard-easy effect was abolished, but with absolute difference a hard-easy effect was still observed. In a second study, we explicitly presented participants with the absolute difference information for each pair of countries to investigate whether or not this would abolish the hard-easy effect. The results are discussed with respect to ecological and cognitive-bias accounts of calibration performance.

Judging confidence influences decision processing

William M Petrusic1 and Joseph V Baranski2

1. Carleton University, Ottawa

2. DCIEM, Toronto

We examined the effects of rendering confidence judgments on the properties of the decision process in sensory-perceptual and general knowledge tasks. The requirement of confidence judgments substantially increases decisional response times and, generally, the increase is inversely related to the difficulty of the judgement. When confidence is rendered, discriminative accuracy may also be impaired; although these effects are small in absolute terms, the relative decreases in accuracy are large and are an inverse function of decision difficulty. Finally, in a sensory detection task, we found that decision times increased with the number of confidence categories used (two, four, or six) and the increases in decision time were the largest when detection was easiest. Taken together, these experiments implicate both decisional and post-decisional loci for the basis of confidence and they permit rejection of the view that the confidence-based increases in decision time arise as a consequence of a slowed rate of evidence accumulation due to dual-task processing.

Explaining extreme overconfidence on interval questions

Jack B Soll1 and Joshua Klayman2

1. INSEAD, Fontainebleau, France

2. University of Chicago, USA

Overconfidence depends significantly on how questions are asked. For example, people are only mildly overconfident on binary choice questions, but they are tremendously overconfident on interval questions (Klayman, Soll, González-Vallejo & Barlas, 1999). In our current research, we develop a random error model for intervals to help explain this extreme overconfidence. The basic idea is that if people are unreliable in how they set interval limits, that should translate to greater uncertainty and wider intervals. People will be overconfident to the extent that they fail to account for variation in the way that they set the limits. In addition, we hypothesize that interval questions are more susceptible to confirmation biases, because the alternative hypothesis is not as salient as in binary choice questions. We use models, simulations, and experiments to illustrate and test these propositions.

Subjective probability in the assessment of threat: Comparing expert vs. novice use of inconclusive information

Joseph V Baranski1 and William M Petrusic2

1. DCIEM, Toronto

2. Carleton University, Ottawa

This study examined how people use inconclusive information when forming subjective probability assessments using a medium fidelity Naval threat assessment simulation. In the present context, inconclusive information refers to data that is relevant but does not clearly support a decision alternative. On each of 36 trials, subjects interrogated 10 pieces of information (e.g., speed, direction, bearing) about 'targets' in a radar space. The amount of hostile (n(H)), peaceful (n(P)), and 'inconclusive (n(I)) information was factorially varied across targets. For novices (i.e., civilian university students), inconclusive information "dilutes" threat assessments; i.e., when in doubt, civilians conservatively err on the side of "friendly". For experts (i.e., senior OPs room officers), inconclusive information is ignored, except when the difference between n(H) and n(P) is small; i.e., when in doubt, military experts conservatively err on the side of "foe". Subjective probabilities in the assessment of threat are best fit by a model that includes a component based on the 'balance of evidence' [i.e., n(H) - n(P)] and a component based on the scaling of doubtful evidence [i.e., n(I)].

(End of symposium)

Session C - Friday 21 July 11-1 pm

The spatial tuning of human auditory perceptual channels under binaural and monaural conditions

Susan E Boehnke and Dennis P Phillips

Dalhousie University, Halifax

Human sound localization is acute for frontal locations and poorer in the lateral hemifields. The tuning of the perceptual channels for auditory space that underlie this acuity has not previously been examined. We have measured this tuning by exploiting a "between-channel" gap-detection paradigm in which detection thresholds for a silent interval (gap) were obtained when the leading and trailing noise markers of the gap were disposed at various absolute locations and separations throughout 360 degrees of azimuth, under binaural and monaural conditions. The binaural results indicate there are two, broadly-tuned hemifield channels whose medial borders overlap through the midline. The monaural results were systematically different, confirming the spatial channels were due to binaural processes. These results are compatible with data on the spatial receptive fields of cortical auditory neurons, and they provide a framework which may explain the localization acuity of human listeners, as well as some spatial masking and attentional phenomena.

The effect of spectral slope on infants' discrimination of timbre

C D Tsang and L J Trainor

McMaster University

Different speech sounds, voices, and musical instruments differ in spectral slope (the rate at which intensity falls off with frequency), and their spectral slopes cluster around -6 dB/octave. If spectral slope is important in timbre perception, enhanced sensitivity would be expected for spectral slopes near -6 dB/octave in comparison to spectral slopes very different from this value. Spectral slope discrimination was tested in 8-month-old infants across a range of spectral slopes (-16 vs. -10; -10 vs. -4; -3 vs. +3; +4 vs. +10; and +4 vs. +16 dB/octave) using a conditionedhead turn procedure. Infants only showed discrimination between -10 and -4 dB/octave. This suggests that the auditory system is tuned to sounds in the range of real-world spectral slopes very early in life.

Hierarchical processing in the auditory cortex? fMRI mapping of responses to spectral and temporal complexity

D A Hall1, I S Johnsrude2, M S Gonçalves1, M P Haggard1, A R Palmer1, A Q Summerfield1, M A Akeroyd1 and R S J Frackowiak2

1. MRC Institute of Hearing Research, Nottingham

2. Wellcome Department of Cognitive Neurology, London

The present study tested the hypothesis that processing in primary and secondary auditory areas is hierarchical: complex acoustic stimuli should activate secondary auditory regions more strongly than do simple stimuli. We used a 2x2 factorial design created by crossing spectral complexity (harmonic complex with F0=186 Hz vs single tone at 500 Hz) and temporal complexity (frequency modulated at 5 Hz vs static signal). Signals were presented diotically through electrostatic headphones to 6 normally-hearing volunteers. Functional and structural images were acquired using a 2T Magnetom VISION MRI system. The 4 conditions were analysed using a fixed-effects multi-subject model (SPM99).

There were significant main effects of both spectral and temporal complexity (P<0.05). Both modulation vs static and complex vs single tone contrasts produced activation bilaterally in the primary auditory area and immediately adjacent secondary auditory areas: activation was largely additive (no significant areas of interaction). These main effects broadly co-localised, but the region responsive to temporal complexity was the more extensive. Responses to temporal complexity alone were found in additional areas bilaterally in the superior temporal sulcus (STS) and in a region anterior to the primary auditory area on the supratemporal plane (AA). Additivity may suggest an increasing auditory response to complexity. However, support for hierarchical processing is limited because effects of spectral and temporal complexity occurred in both primary and secondary areas. The specificity of the response in STS and AA to temporal complexity merits further investigation.

Subsystems in the human auditory cortex - evidence from functional neuroimaging studies

Sophie K Scott1, Stuart Rosen2, Catrin Blank3 and Richard J S Wise3 (Introduced by James Blair)

1. Institute of Cognitive Neuroscience, University College London

2. Department of Phonetics, University College London

3. MRC Cyclotron Unit, Hammersmith Hospital

Work in visual neuroscience has identified at least two possible functional subsystems in visual processing, a ventral 'what' pathway passing from visual cortex to the ventral temporal lobes and a dorsal 'where/how' pathway passing from visual cortex to the parietal lobes. Work in non-human primates has revealed analogous pathways in the auditory system. Using speech stimuli of varying intelligibility but equivalent structural complexity, and functional neuroimaging (PET) we demonstrate a stream of processing in the human left temporal lobe associated with the intelligibility of speech The left superior temporal sulcus responds to the presence of phonetic information, but its anterior part only responds if the stimulus is also intelligible. This novel observation demonstrates a left-lateralised anterior temporal pathway for speech comprehension. We propose that this forms part of a verbal 'what' pathway in human auditory perception. The right temporal lobe, in contrast, showed selective activation for stimuli that contained dynamic pitch variation, regardless of intelligibility.

Multisensory coding: Implications of findings on tactile illusions

Susanna Millar and Zainab Al-Attar

University of Oxford

The similarity of errors in touch to 'optical' illusions for T-shapes and Müller-Lyer figures presents a paradox that is not resolved by modality-specific nor by cognitive mediation theories. Tactile shape perception depends on inputs from touch, movement and body posture cues, in the absence of vision. Nevertheless, the vertical line is overestimated in inverted T-shapes and underestimated in rotated T-shapes, as in vision. Similarly, the shaft of figures with converging fins is perceived as smaller by touch than in divergent fin figures. We argue that the illusions can be explained by disparities in spatial reference cues from all sources, which normally converge (coincide) in accurate perception. We show that vertical errors in T - shapes are predicted by the locations of the additional junction point on segmented lines, and that the illusion is reduced significantly by two-handed scanning relative to an external frame and to egocentric cues. The Müller-Lyer illusion is explained by disparities between cues to shape and length from the fin and shaft elements within the shapes. Instructions to ignore the fins had no effect in vision in the presence of allocentric cues, but no movements relative to egocentric cues. The tactual illusion was almost eliminated by instructions to relate scannin movements to egocentric cues and to ignore the fins. The hypothesis, that length illusions in touch and vision depend on disparities in current spatial reference information, accounts for similarities and differences between the visual and tactual illusions by assuming that accurate perception in both modalities depends on congruent reference cues from all sources.

How does the brain solve the crossmodal binding problem? Insights from fMRI studies of audio-visual and visuo-tactile integration.

Gemma Calvert

University of Oxford

The integration of information from different sensory channels markedly enhances the detection and identification of objects and events in the environment. Semantically and/or spatially congruent multisensory inputs speed discriminations and improve reaction times compared to the response to any single modality in isolation. Incongruent inputs have the opposite effect, hampering classification and slowing responsiveness so that erroneous binding is prevented. These behavioural features of crossmodal processing appear to have parallels in the response properties of multisensory integrative cells in the superior colliculus in lower mammals. Spatially concordant multisensory inputs result in a substantial, often multiplicative (i.e. greater than the sum of the individual inputs) response enhancement; discordant inputs can produce a profound response depression. At present, it is not known whether a similar method of intersensory integration onto multisensory neurons applies to humans, or for the integration of complex inputs relating to stimulus identity, as well as stimulus location. This question has now been addressed in a series of experiments using fMRI. In the first series of experiments, subjects were exposed to alternating epochs of congruent and incongruent audio-visual (speech and non-speech) inputs and to each modality in isolation. In a second series of experiments, subjects were scanned whilst presented with spatially concordant or discordant visuo-tactile inputs (vibro-tactile stimulation and LED's) whilst attention to one or modality was manipulated by the response instructions. The results of these studies provide clear neurophysiological evidence of crossmodal binding by convergence onto multisensory cells in human heteromodal cortex. They further suggest that response enhancement and depression may be a general property of multisensory integration operating at different levels of the neuroaxis and irrespective of the purpose for which sensory inputs are combined.

Session A - Friday 21 July 2-3.20 pm

Symposium: "Parietal lobe in vision and visuomotor control"

Organisers: Melvyn A Goodale and A David Milner

Frames of reference and timing in the visual control of skilled actions.

Melvyn A Goodale and Yaoping Hu

MRC Group on Action and Perception, University of Western Ontario

We recently examined the frames of reference used by normal subjects to scale their grasp when picking up objects of different sizes. When faced with a target object and a distractor that was either larger or smaller than the target, it did not matter which distractor the target was paired with: the grasp was scaled to the real size of the target object and was not affected by the presence of the other object. When subjects picked up the target after a 5-s delay during and after which the target was no longer visible, a significantly larger grip aperture was observed when the target had been paired with a smaller object, than when the same target had been paired with a larger object. Exactly the same effect was observed when subjects simply gave a ‘perceptual’ estimate of the target’s size – even when the target was visible. These results suggest that normal (real-time) visuomotor control relies on real-world metrics whereas delayed grasping utilizes the same relative (scene-based) metrics used by conscious perception. This result converges nicely with previous experiments on the visual-form agnosic patient D.F. who shows completely normal scaling of her grasp in real time but loses all scaling when a 2 s delay is introduced between viewing the object and making the movement. D.F. presumably can use visuomotor networks in her intact parietal cortex to program real-time actions but cannot use stored perceptual representations to drive a grasp ‘off-line’.

Two routes from vision to action: Effects of bilateral posterior parietal damage.

A David Milner1 and H Chris Dijkerman2

1. University of Durham

2. University of Utrecht, Netherlands

Goodale and colleagues showed some years ago that the visual-form agnosic patient D.F. was quite unable to retain the visual dimensions of objects when required to withhold response for a few seconds in order then to make a 'pantomimed' response in the absence of the object. We later found that D.F. even has difficulties in directing her eyes or hand to the location of an LED after a few seconds of delay. These observations fit with the idea that making a pantomimed response requires that explicit perceptual representations be used to guide action ‘off-line’, a capacity that D.F. no longer possesses because of her brain damage. We have recently studied two patients with Balint-Holmes syndrome, resulting from bilateral posterior parietal damage, to seek complementary evidence for this 'indirect' route from vision to action. These patients show 'optic ataxia'; that is they are impaired in the direct visuomotor control of actions such as reaching and grasping. We have found, however, that they show paradoxical improvements in their visuomotor control when making pantomimed rather than direct responses to target objects, in both the spatial and size domains. This is consistent with the idea that they can use a relatively well-preserved 'perceptual' route to action. We also found evidence that one of the patients spontaneously developed a 'perceptual' strategy to overcome her visuomotor difficulties.

How does the posterior parietal cortex control visually guided actions?

Michel Desmurget.

Psychophysique et Neuropsychologie Espace et Action, INSERM Unité 94, Bron, France.

Numerous studies have indicated that the posterior parietal cortex (PPC) is critically involved in the execution of visually-directed movements. Although PPC involvement in planning processes has been emphasized in the past, it was often suggested that this area may also be involved in on-line movement control. In this talk I shall provide direct evidence supporting this view. Based on recent studies involving neuroimaging, transcranial magnetic stimulation, and brain damaged patients, I shall defend the idea that PPC (1) generates an internal representation of the instantaneous hand location by integrating the motor outflow and sensory inflow (forward model), (2) computes a dynamic motor error by comparing the estimated state of the hand location with the location of the target.

Visually guided grasping produces fMRI activation in human anterior intraparietal sulcus.

Jody C Culham1, Rieko Osu1, A David Milner2, Joseph S Gati1, Ravi S Menon1, and Melvyn A Goodale1.

1. MRC Group on Action and Perception, University of Western Ontario.

2. University of Durham

We used functional magnetic resonance imaging (fMRI) to examine the human neural substrates of visually-guided grasping. Within the relatively wide bore of a 4 Tesla magnet, subjects directly viewed (without mirrors) a rotating drum, which presented rectangular plexiglas blocks of varying length and orientation. On each trial, a single target was illuminated and the subject reached towards the object and grasped its edges with a precision grip (using the index finger and thumb). In a control condition, subjects reached and touched but did not grasp the target object. Event-related single trials took advantage of the haemodynamic delay to dissociate true grasping activation from motion artifacts. Grasping produced higher activation levels than reaching in a region of anterior intraparietal (AIP) cortex. This area is a likely human homologue of monkey AIP, an area containing neurons which fire in preparation for grasping. In addition, several other parietal and occipito-temporal regions showed grasp-related activation. Thus we have found a reliable paradigm which allows us to identify human AIP and other grasping-related regions. This will allow us to conduct future experiments comparing grasping to other tasks such as object perception (with no motor component), grasping imagery, and delayed grasping.

Session B - Friday 21 July 2-3.20 pm

Symposium: "New research on human motivations for alcohol use"

Organiser: S H Stewart

Adolescents' motives for cigarette smoking and marijuana use: Factor structure and relations to motives for alcohol use

N Comeau, S H Stewart, P Loba, E Rhyno and H L Loughlin

Dalhousie University, Halifax

The present study was designed to test the utility of extending Cooper's (1994) 2 x 2 (valence x source) drinking motives model to adolescents' use of cigarettes and marijuana, and to examine the degree of correspondence of specific substance use motives across three different drugs (i.e., alcohol, cigarettes, and marijuana). Of a sample of 508 adolescents (239 females, 269 males; M age = 15.1 years), 312 (61.4%) reported alcohol use (M = 6.7 drinks/week), 192 (37.8%) reported smoking (M = 40.3 cigarettes/week), and 154 (30.3%) reported use of marijuana (M = 6.1 marijuana cigarettes/week). Factor analytic and multiple regression results support Cooper's four-factor model of drinking motives, and suggest that this model may also be applied to adolescents' reasons for smoking and marijuana use. The results further suggest that there are some similarities in adolescents' specific motives for substance use across different drugs, particularly between motives for alcohol and marijuana.

Relations of drinking motives to promotion- and prevention-focus in other behavioral domains

S H Stewart1, C J Roney2, D R Lehman3, L Chung3, N Lawson3 and A Mueller3

1. Dalhousie University, Halifax

2. Kings College, Ontario

3. University of British Columbia

We examined whether positive (Enhancement; Social) and negative (Coping; Conformity) drinking motives (Cooper, 1994) relate to promotion- or prevention-focused self-regulation (Higgins, 1997) in the domains of social interaction and academic achievement. In four multiple regressions, we used 158 undergraduates' drinking motives scores to predict levels of promotion- and prevention-focus in social and academic domains, respectively. Positive and negative emotional drinking motives in Cooper's model (cf. Cox & Klinger, 1990) related more broadly to promotion- and prevention-focus, respectively, in domains other than drinking. Those who drink to increase positive emotions report a promotion-focus in their orientation toward social interactions and academic achievement (striving to increase affiliation/success). Those who drink to avoid negative emotions report a prevention-focus in their orientations toward social interactions and academic achievement (striving to avoid rejection/failure). Results suggest that people differ in their general proclivities to either approach pleasure or avoid pain across a variety of behavioral domains.

Relations between four personality risk factors for alcohol abuse and reinforcement-specific motives for drinking in non-alcoholic and substance abusing samples

P J Conrod1, P Woicik1, S H Stewart2 and R O Pihl3

1. State University of New York at Stony Brook, USA

2. Dalhousie University, Halifax

3. McGill University

Four personality characteristics (anxiety sensitivity; introversion/hopelessness; sensation seeking; impulsivity) discriminate between subtypes of substance abusers who demonstrate patterns of dependence on drugs that possess specific reinforcing effects (Conrod et al., in press). Since alcohol possesses a number of different reinforcing properties, we hypothesized that these personality risk factors should also be associated with self-reported reasons for alcohol use that are reinforcement-specific. 300 substance abusing women and 450 male and female undergraduates completed the Drug Abuse Subtyping Scale (Woicik & Conrod, 1999) and Drinking Motives Questionnaire (Cooper et al., 1992). Results were almost identical across the clinical and non-clinical samples. Anxiety sensitivity and hopelessness were specifically related to drinking motives involving coping with negative emotion. Sensation seeking was associated with drinking for social and enhancement effects. Impulsivity was associated with drinking for both positive and negative reinforcement. Results are discussed within the context of a four-factor model of drug abuse vulnerability.

Motivational patterns as predictors of alcohol use and the likelihood of change

W M Cox

University of Wales, Bangor

In a series of three studies, the Motivational Structure Questionnaire (MSQ) was used to predict alcohol consumption and motivation for change. The MSQ identifies respondents' salient concerns in various life areas, and the characteristic manner in which they strive for goals to resolve their concerns (i.e., their motivational structure). Participants in the three studies were college students or patients with alcohol abuse. Across the three studies, respondents with high scores on an "adaptive" motivational factor felt committed to goal pursuits from which they expected to derive emotional satisfaction, and they felt optimistic about succeeding. The adaptive factor was a significant negative predictor of (a) the amount of alcohol that the college students habitually consumed, (b) alcohol abusers' denial of a problem, and (c) the amount of alcohol that alcohol abusers drank one year following treatment. The adaptive factor was also a significant positive predictor of alcohol abusers' readiness for change.

(End of symposium)

Session C - Friday 21 July 2-3.20 pm

PCA network models of memory and perceptual processes: A matter of style

John R Vokey1, Jason M Tangen2 and Kevin Eva2

1. University of Lethbridge

2. McMaster University

The ability of people and animals to demonstrate tacit sensitivity to the structure of the environment around them is explored in terms of a simple PCA network model applied to pixel-maps of natural and artificial images. The eigenvectors obtained from the singular value decomposition of sets of these pixel-maps provide for descriptions of the stimuli in terms of visual "macro-features". These macro-features provide a simple basis not only for recognising previously-experienced images, but for the successful discrimination of novel images into various natural and artificial structural categories. A summary of simulations of the performance in various focal domains of humans (e.g., face categorisation/recognition, artificial grammar learning, EKG diagnoses), pigeons (natural categories, artists' paintings), and chimpanzees (kin recognition) is provided. The results suggest both that the eigen-decomposition is a necessary first-step, and that the discrimination is best seen as a judgement of the "style" of family-resemblance structures.

Categorization and the "Ratio Rule

A J Wills (Introduced by I P L McLaren)

Cambridge University

Many modern theories of categorization are based on the production of magnitude terms which represent the subjective level of evidence for a presented stimulus being a member of a particular category. Almost universal amongst such theories is the assumption that magnitude terms may be translated into response probabilities by the Ratio Rule (a.k.a. the Choice Axiom - Luce, 1959). Through the presentation of a number of categorization experiments, a case is made that the Ratio Rule is an inappropriate formulation. It is suggested that theorists should employ a different type of Thurstonian choice process.

Luce, R. D. (1959). Individual Choice Behavior. New York: John Wiley & Sons

Relation between perceptual and informational learning of family resemblance structures

Lee R Brooks and Sam D Hannah

McMaster University

In the learning of both medical and artificial materials, explicit rules commonly function to name the objects of perceptual learning rather than to give sufficient conditions for identification. Subjects in experiments with artificial materials search for informational commonalities even when the identification task is incidental to the apparent main purpose of the experiment. However, the rules so generated do not describe their dependence on particular perceptual manifestations. A comparable phenomenon is observed in learning to apply medical rules. The standard, authoritative rules are clearly useful in learning to identify medical disorders in visual areas (ECG, patient appearance, dermatology), but they do not describe the dependence on particular perceptual manifestations that is acquired during the course of learning. The influence on these phenomena of family resemblance structure, featural ambiguity and confidence in rule application will be discussed.

The under-weighting of implicitly generated alternatives

Kevin W Eva and Lee R Brooks

McMaster University

A highly robust finding in the study of subjective probability judgments is that the rated likelihood of a hypothesis depends on the other hypotheses suggested (called the unpacking effect by Tversky and Koehler, 1994). For example, the rated probability that a randomly selected person might die of cancer tends to be greater when cancer is considered by itself than when presented in a list of potential diagnoses. Our previous work has shown, somewhat counter-intuitively, that the alternative diagnoses that have the greatest influence on the probability rating of a focal diagnosis are those that were most likely to have been considered implicitly (Eva, Brooks, Cunnington, and Norman, Submitted). This result suggests that participants under-appreciated diagnostic alternatives that they themselves generated relative to those that were explicitly presented. The current study was designed to demonstrate this same result for alternatives the participants themselves claim to have actually considered. We used clinical cases previously shown to have two highly likely and roughly equiprobable diagnoses. Second year medical students viewed either one or both of these diagnoses and were asked to rate them. They were then asked whether they had been considering a diagnosis that was not presented and if so what it was. A week later they were shown the original cases and asked to rate the original alternative(s) together with the alternative they had generated. The unpacking effect was found even when the diagnostic alternative that is unpacked is one that they claim to have considered while originally viewing that case. That is, they rated the originally presented diagnosis as less probable when the alternative they had claimed to be considering implicitly was explicitly presented. Furthermore, patient management strategies, such as requesting diagnostic tests, were affected in addition to pure probability ratings. Possible reasons for this under-appreciation of self-generated diagnoses will be discussed.