1 The identification of pheromonal components in male mouse urine which initiate puberty acceleration in female mice
Heather M Schellinck1, Allison Clarke1, Richard E Brown1 and Michael Wilkinson2
1. Department of Psychology’ Dalhousie University, Halifax
2. Department of Obstetrics and Gynecology, Dalhousie University, Halifax
The pheromonal components of male mouse urine which accelerate puberty in female mice have not been clearly identified. To clarify this issue, we examined the effect of a number of compounds on puberty acceleration and neural activation of the accessory olfactory bulb (AOB). During postnatal days 21-28, juvenile female mice, were exposed to male urine, major urinary proteins (MUPs), MUPs+ligands attached (2-sec-butyl-dihydrothiazole:SBT and 2,3-dehydro-exo-brevicomin:DHB ), SBT/DHB alone or saline. Puberty acceleration occurred in mice exposed to male urine or MUPS+ligands but not in mice exposed to SBT/DHB alone or MUPs alone or saline. Immediate early gene expression (c-fos) was significantly higher in the AOB of mice exposed to male urine orMUPS+ligands than in those mice exposed to saline, SBT/DHB or MUPS alone. These results indicate that exposure to specific components of male mouse urine not only induces puberty acceleration but also activates the neural pathway which initiates such an effect.
2 Behavioural and cognitive response to chronic ritalin administration during prepubertal development in mice.
M P McFadyen1, N Carrey2 and R E Brown1
1. Department of Psychology1, Dalhousie University, Halifax
2. Department of Psychiatry2, Dalhousie University, Halifax
Methylphenidate hydrochloride (Ritalin, MPH) is frequently prescribed as a treatment for children with attention deficit hyperactivity disorder (ADHD); however, little research has been conducted to determine its potential long-term neurobehavioural effects. We assessed the effects of chronic MPH administration on male CD-1 mice treated from 26-32 days of age. When tested at 33 days of age in the open field and elevated plus maze, there were no significant differences in spontaneous locomotion, exploratory, fear-, or anxiety-related behaviours. When tested from 34 to 37 days of age in the Morris water maze, we found no significant effects of any dose of MPH on learning or memory in the mice. While it is difficult to extrapolate directly from these results to clinical effects in humans, our results indicate that pre-exposure to MPH late in the developmental period of mice does not appear to alter later behaviour. Studies that probe further into the potential effects of MPH administration during development are needed to replicate these findings, and to examine other contributing factors, such as developmental stage or duration of MPH administration.
3 Effects of chronic and acute ritalin treatment on mouse pup development
M R Penner1, M McFadyen1, R E Brown1 and N Carrey2
1. Department of Psychology1, Dalhousie University, Halifax
2. Department of Psychiatry2, Dalhousie University, Halifax
Methylphenidate hydrochloride (MPH, Ritalin) is a psychostimulant used to treat children with Attention Deficit Hyperactivity Disorder (ADHD). Despite its widespread use, the consequences of MPH treatment is not well documented in the developing nervous system. This study examined the effects of chronic and acute MPH treatment on isolation-induced ultrasonic vocalizations (UVs), spontaneous locomotor activity, and other aspects of neuromotor development in 3-11 day old outbred CD-1 mouse pups. In experiment 1, pups received daily injections (sc) of either 5mg/kg or 20mg/kg of MPH, saline, or no injection. Testing took place on postnatal days 3, 5, 7, 9, and 11. MPH treatment had no significant effect on body weight, neuromotor development or the number or duration of UVs emitted. However, both doses of MPH resulted in significant increases in locomotor activity. In experiment 2, pups were given a single dose of MPH (5mg/kg or 20mg/kg) or an equivalent amount of saline on one of PND 5, 7, 9, or 11. This acute MPH treatment increased locomotor activity but had no significant effects on neuromotor development or isolation-induced UVs. These results suggest that neither acute nor chronic MPH treatment have short-term effects on the developing nervous system.
4 Synaptic release of norepinephrine from the locus coeruleus in vivo
increases epsp slope in putative late-phase potentiation in the dentate gyrus.
R A M Brown, S G Walling and C W Harley
Memorial University of Newfoundland
Norepinephrine (NE) modulates learning in a variety of tasks. Tetanic long-term potentiation (tLTP) is a model of cellular learning that has distinct early and late phases. In dentate gyrus tLTP facilitates the population spike (Pop spike), a measure of postsynaptic cell firing, and the excitatory postsynaptic potential (EPSP) slope, a measure of synaptic drive. NE depletion, or blockade of =DF-adrenoreceptors, reduces the probability that tLTP will be elicited in the dentate gyrus. Synaptically released NE can induce an LTP independent of tetanic stimulation in vivo (NE-LTP). Previous work with synaptically released NE has determined that NE-LTP potentiates the Pop spike for at least 30 minutes, but NE-LTP has not been investigated at later time-points. Here we report that NE-LTP can last at least 3 hours, a time-point consistent with late-phase tLTP in in vitro preparations. In addition, Pop spikes that are potentiated at 3 hours are associated with increased EPSP slopes.
5 Dopamine efflux in the rat nucleus accumbens during exploratory foraging on a radial-arm maze
Soyon Ahn, Stan B Floresco and Anthony G Phillips.
University of British Columbia
Searching for food in an unpredictable environment on a radial arm maze is dependent on a neural circuit linking the hippocampus and the nucleus accumbens (NAc; Floresco et al., 1997). In addition, mesoaccumbens dopamine (DA) transmission plays an essential role in this behavior (Floresco and Phillips, 1999). In the present study, DA efflux in the NAc was measured in vivo with brain microdialysis while rats searched for 4 pieces of food placed at random on 4 of the 8 arms of a radial maze. Rats that were permitted to explore the maze for food displayed a significant increase (~50%) in DA extracellular levels in the NAc, which remained elevated during a 5 min period of exploration, and for a another 5 min while the rats were confined to the center of the maze. These data, in addition to our previous findings suggest that increased mesoaccumbens DA transmission is an important component of exploratory behaviors which are mediated by hippocampal-ventral striatal circuitries. Future experiments will asses the effect of reward omission on the changes in mesoaccumbens DA release during exploratory foraging.
6 The influence of lesions of the nucleus accumbens and lateral hypothalamus on anxiety and reactivity in the male rat.
Tod E Kippin, Veneta Sotiropoulos and James G Pfaus
Center for Studies in Behavioral Neurobiology, Concordia University, Montreal
The role of nucleus accumbens and lateral hypothalamus in emotional behavior was examined by assessing the effects of lesions of these structures on the elevated-plus maze and the resistance-to-capture tests. Male Long-Evans rats received either 1.0 ul injections of NMDA (20 ug/ul) or vehicle into the nucleus accumbens or lateral hypothalamus. Rats were allowed to recover until feeding and body weight stablized then were subjected to behavioral tests. Rats receiving either lesion displayed more resistance to capture than did controls with the nucleus accumbens lesioned rats displaying the highest reactivity. On the elevated-plus maze test, rats receiving either lesion spent more time in the closed arms and made fewer crosses than did controls. These results indicate that disruption of the nucleus accumbens or lateral hypothalamus alter the expression of emotional behaviors.
7 Conditioning of Interictal Behaviors by Amygdala Kindling
Steven J Barnes, John P J Pinel, Lee H Francis and Gagan S Wig
University of British Columbia
We assessed the ability of contextual conditional stimuli that are normally present during the course of kindling to modulate interictal behavior. Rats received 53 stimulations to the left basolateral amygdala in one context (CS+) and 53 sham stimulations (lead was attached but no current was delivered) in another context (CS-), quasirandomly over 53 days. We observed 2 kinds of conditional effects. First, after several stimulations, less ambulatory activity, more freezing, and less rearing reliably occurred in the CS+ context than in the CS- context. Second, after 45 stimulations, all of the rats chose the CS- context over the CS+ context in a conditioned place preference test. Thus, the conditional effects on interictal behavior appear to be an integral, but unrecognized component, of the kindling phenomenon.
8 The basolateral amygdala (BLA) is involved in sensory and not motivational representations of outcomes
P J Blundell1 and A S Killcross2
1. University of York
2. Cardiff University
We have shown previously that rats with BLA lesions fail to show a benefit of training with differential outcomes in discrimination learning, and proposed that the BLA modulates the development of representations of the sensory properties of reinforcers (US). We have used a Pavlovian-instrumental transfer task to confirm that BLA-lesioned rats can access the affective, but not sensory, properties of USs. Rats were trained to press one lever to obtain sucrose solution, and another to obtain a food pellet. In separate sessions they also received Pavlovian conditioning with one CS predicting pellet delivery, and a second CS, sucrose. In an extinction test, lever pressing was measured in the presence of each CS, and in their absence. Sham-lesioned rats made more lever presses during the CSs than in their absence, and this effect was more marked on the lever that had produced the reinforcer predicted by each CS. BLA-lesioned rats also responded more during the CSs, indicating that the arousing properties of the appetitive CSs were intact. However, this increase in responding was not greater on the lever matched to the CS outcome, indicating the failure of BLA-lesioned rats to represent the specific characteristics of the US.
9 Examination of conditional control of place fields by black/white discrimination
J M Barry, G M Martin, C W Harley and D T Laidley
Departments. of Biopsychology and Psychology, Memorial University of Newfoundland
We examined the influence of discrimination learning on the formation of cognitive maps in the rat. We determined whether or not learners of a black/white discrimination would treat goal and delay areas of the test apparatus as different places conditional upon their performance. Lawrence and Hommel (1961) showed that rats would learn a black/white discrimination over a delay when the goal boxes were of different color. The rats could not learn a black/white discrimination over a delay when the goal boxes were of the same color. These results suggested that the experience of reward and non-reward in different environments improves the learning of black/white discrimination. Rats did not demonstrate learning of the discrimination unless a delay at the choice point was imposed. Preliminary data from hippocampal pyramidal cells suggests that rats which exhibit the same cognitive maps in reward and non-reward environments show poor discrimination learning.
10 Wheel running, feeding, and weight in weanling and adult male rats
V M Afonso. and R Eikelboom
Wilfrid Laurier University
The effects of running wheel introduction on running, feeding, and body weight were investigated in 21- (weanling, n=21) and 50- (adult, n=16) day old male Sprague-Dawley rats. Body weight was measured daily while feeding, and wheel running was measured every 5 sec. After a 4 day baseline, 12 weanling and 8 adult rats were given access to wheels for 32 days. Wheel running was similar in weanlings and adults, averaging over 5000 wheel turns a day by the end of the experiment. Only in adults did wheel access suppress feeding for the initial 5 days. After 7 days, wheel access had no effect on feeding. At both ages, wheel access initially suppressed weight by about 5%, but only in adults was the weight suppression maintained throughout the experiment. This suggests that the rat's age at wheel introduction is a critical variable in the wheel's effects on feeding and weight.
11 How well do boys do? Colour and pattern recognition by male bumblebees
Dana Church and C M S Plowright
University of Ottawa
Male bumblebees have been well-studied in such areas as mating and territoriality, however the area of male bumblebee cognition has been neglected. This study explored colour and pattern discrimination by male bumblebees: two critical abilities as males must forage for themselves once they leave the colony. Males were removed from a colony of Bombus impatiens and trained to forage from: (a) blue and yellow artificial flowers, where one colour was rewarding and the other colour unrewarding; or (b) artificial flowers bearing one of two patterns, one rewarding and the other unrewarding. Results from testing sessions (where no flowers were rewarding) showed that males could discriminate well between yellow and blue flowers but could not discriminate between the two patterns.
12 Why do orally consumed calories fail to condition preferences for relatively unacceptable tastes?
Catherine A Forestell and Vincent M LoLordo
Dalhousie University, Halifax
A reverse-order differential conditioning procedure was employed to condition taste preferences in rats. During training, one taste solution (CS+) was presented 3 min after glucose, and another taste solution (CS-) was presented 3 min after water. For one group of animals (Group Sacc), the CS solutions were sucrose octaacetate (SOA) and citric acid (CA) mixed in saccharin, whereas for another group (Group No Sacc) SOA and CA were mixed in water. Subsequently in 2-bottle tests, only Group Sacc expressed a conditioned taste preference for the stimulus paired with glucose. To determine whether low consumption of the taste cues during conditioning prevented the No Sacc Group from acquiring a conditioned taste preference, a second experiment was conducted in which training consumption of the taste cues in Group Sacc was yoked to that of Group No Sacc. Results will be discussed.
13 Representations of serial reward patterns in the T-maze by rats
J S Cohen, Ann-Marie Simpson, Kim Westlake and Michelle Pepin
University of Windsor
We report two experiments in which rats acquired two serial reward patterns over three massed trials in a T-maze: RRN and RNR. One group of rats experienced each pattern paired to a specific arm, e.g,, RNR on left arm and RRN on right arm (Arm Associated group) while the other group experienced each pattern equall on each arm (Nonassociated group). Rats were first run on the same pattern over succesive three-trial sequences within a session and then received each sequence in a random order within each session. Following these two training phases, rats received test phases with free-choice trials either on the last two or on all three trails within each sequence. Running speeds during the first forced-choice training phase indicated that all rats learned to anticipate the N-trial within each sequence by running slower on that trial (trial tracking). Running speeds on the second training phase, showed that the Arm Associated rats were able to develop accurate track trialing within each pattern but the Nonassociated group was only able to track the RRN pattern. How well the latter rats were able to accomplish this partial trial tracking was based on the general shaping procedures used prior to introduction of the three-trial sequences (Experiments 1 vs 2). Of greater interest were the results of the partial and complete free-choice tests. The pattern of arm alternations indicated that the Arm Associated group may have developed an ordinal position representation of reward sequences to each arm while the Nonassociated group were only able to use a rule based representation in determining which arm to select on its last choice.
14 Use of geometric and featural cues in a touch-screen environment
Debbie M Kelly and Marcia L Spetch
University of Alberta
Many spatial cues may be used to develop a spatial representation of one’s environment. In previous studies, pigeons showed use of geometric and featural information to find food located in one corner of a rectangular enclosure. We investigated pigeons’ relative use of geometric and featural cues in a 2-dimensional environment. Using a touch screen mounted to a computer monitor, pigeons were presented with images of a rectangular environment. One corner of the rectangle was defined as the target area. Only pecks directed to the target corner were reinforced. The rectangle’s orientation on the screen varied across trials. In a Geometric condition, corners could be distinguished only by the rectangle’s metric properties, whereas in a Feature condition, distinct features were provided in each corner so that either geometric or featural information could be used. The pigeons were tested with transformations altering the metric or featural properties of the environment.
15 One or multiple mechanisms underlying orientation-invariant object identification?
K E Jaskie and P A McMullen
Dalhousie University, Halifax
Is there a single or multiple mechanisms by which rotated objects are identified independently of orientation? Orientation-invariant identification of objects occurs at the basic and superordinate levels of identity, as well as after practice with naming objects. The underlying mechanism(s) might be based upon the extraction of orientation-invariant features, and/or the global attributes (requiring intact spatial arrangement of the features) of objects. We presented four blocks of feature-scrambled objects followed by a block of intact, rotated objects, in a word-picture verification task that allowed for control of the level at which the objects were identified. If a single, feature-based mechanism underlies orientation-invariant identification then, superordinate- and basic-level identification of scrambled objects should be faster than subordinate-level. Furthermore, practice at identifying scrambled objects should diminish orientation effects in the final block. A global-based mechanism would not produce these effects. Alternatively, one effect may be feature-based and the other not.
16 A strong test of the role of axes of elongation and symmetry in rotated object identification.
M E Large, P A McMullen and J Hamm
Dalhousie University, Halifax
Many theorists have postulated that object axes of elongation and/or symmetry play a role in the recognition of objects. Previous studies have demonstrated that these variables influence decisions about the primary axis of objects. However, these studies did not explicitly investigate effects of elongation and symmetry on object identification. In this study, Experiment 1 assessed the effect of the axis elongation (wide vs. tall) on the time to name common objects rotated in the plane. Similarly, Experiment 2 assessed the effect of the axis of symmetry (symmetric vs. asymmetric). The variables of elongation or symmetry did not interact with orientation. In a stronger test of this hypothesis, Experiment 3 orthogonally controlled both of these variables, manipulated the aspect ratio of elongation and used completely symmetric or asymmetric objects. Additional blocks were run to determine whether these variables influenced the diminishment of orientation effects.
17 The neural substrates of biological motion perception: An fMRI study
Philip Servos1, Rieko Osu2 and Mitsuo Kawato3
1. Department of Psychology, Wilfrid Laurier University
2. Kawato Dynamic Brain Project, JST, Kyoto, Japan
3. ATR Human Information Processing Research Labs, Kyoto, Japan
fMRI was used to identify the brain areas related to the perception of biological motion (4T EPI Varian scanner; whole brain; TR=5s). Ten subjects viewed biological motion (a human figure jumping up and down, composed of 22 dots), alternating with a control stimulus created by applying autoregressive models to the biological motion stimulus (such that the dots' speeds and amplitudes were preserved whereas their linking structure was not). By using stimulus bouts of varying lengths the transitions between biological motion and control stimuli were unpredictable. Subjects had to indicate with a button press when these transitions occurred and also performed a task in which they detected short (1s) disturbances within the displays.
To identify shape- and motion-sensitive regions, subjects viewed a series of common objects alternating with band-limited white noise patterns, and linear motion of dots alternating with static dots. Finally, subjects generated imagery of their own arm movements (e.g., scratch back with left hand) alternating with visual imagery of objects (e.g., which larger: airplane or truck). Biological motion specific BOLD signal was found within regions of the lingual and fusiform gyri (generally Brodmann's areas 18, 19, and 37), showing little overlap with object recognition, linear motion, or motion imagery areas.
18 Particular Difficulties with Mirror Images
Jules Davidoff1 and Elizabeth Warrington2
1. Goldsmiths College, University of London
2. National Hospital, London
We report two case studies of patients who presented with dramatic impairments for discriminating between mirror images. The first patient (FH) showed impaired discrimination of mirror images in the context of impaired spatial perception but with very good recognition of objects when presented in conventional views. Despite her generally poor spatial abilities, FH was able to discriminate between inverted and upright versions of stimuli. The marked deficit for mirror images applied to all stimuli though it was more marked for meaningful objects.The second patient (JBA) also demonstrated poor spatial skills and was unable to recognise half of objects even when presented in a conventional view. The ability of JBA to recognise an object was related to her ability to discriminate the object from its mirror image. Paradoxically, an inability to recognise an object was related to good mirror-image discrimination. In consequence, JBA's superior discrimination of mirror-images over inverted/upright discriminations was restricted to meaningful objects.
19 When two is less than one: A new, object-based visual illusion
University of Birmingham
Visual illusions of linear spatial extent (e.g. the Müller-Lyer illusion) have generally been accounted for in terms of contrast or assimilation effects (see Jaeger, 1999 for a summary). While these effects may sometimes depend on the presence or absence of grouping between elements, there has been no demonstration of an illusion which seems to be induced by the nature of grouping and segmentation between target elements. In this paper we present 5 experiments demonstrating the presence of an illusion that is induced by perceptual organisation factors. When elements grouped to form a single perceptual object, distances between target elements were reported as wider than the same physical distances between elements which did not group. The data are consistent with an account based on separate and parallel encoding of spatial relations within single perceptual objects and between multiple perceptual objects, as first suggested by Humphreys and Riddoch (1994; 1995).
Humphreys, G. W., & Riddoch, M. J. (1994). Attention to within-object and between-object spatial representations: Multiple sites for visual selection. Special Issue: The cognitive neuropsychology of attention. Cognitive Neuropsychology, 11(2), 207-241.
Humphreys, G. W., & Riddoch, M. J. (1995). Separate coding of space within and between perceptual objects: Evidence from unilateral visual neglect. Cognitive Neuropsychology, 12(3), 283-311.
Jaeger, T. B. (1999). Assimilation and contrast in geometrical illusions: A theoretical analysis. Perceptual and Motor Skills, 89(1), 249-261.
20 Visuomotor processing of spatial layout in visual form agnosia
Robert McIntosh1, Chris Dijkerman2, Mark Mon-Williams, and David Milner1
University of St Andrews
1. Now at Department of Psychology, University of Durham
2. Now at Psychological Laboratory, University of Utrecht
Patient D.F. suffers from visual form agnosia, a profound deficit of visual recognition consequent on bilateral interruption of the ventral visual pathway. Nonetheless, she can use visual information about a target object (its shape, size and orientation, and its location within the environment) to control her actions. These visuomotor abilities are believed to reflect the intact capacities of the dorsal visual stream which, in D.F., is presumed to operate in relative independence from ventral mechanisms of visual processing. We have now studied D.F.'s performance on two slightly more complex visually guided prehension tasks requiring her to take account of the spatial layout of the workspace. The results of the first experiment showed that D.F. is sensitive to the spatial separation between two objects when opening the thumb and index finger in order to grasp them as a pair. The second experiment demonstrated that D.F. modifies her reaching behaviour, in an anticipatory manner, to take account of potential obstacles to reaching within the workspace. These data extend the known abilities of this much studied patient and suggest that the dorsal stream can incorporate some aspects of spatial layout within its visuomotor control capabilities.
21 Local and global visual processing in normal and impaired face recognition.
Daniel Saumier1 and Martin Arguin2
1. Université de Montréal & Centre de Recherche
2. Institut Universitaire de Gériatrie de Montréal.
Priming studies involving neurologically intact observers and a prosopagnosic patient (AR) investigated the issue of whether the processing of holistic facial information relies on the prior extraction of local facial parts. Neurologically intact subjects show significant priming from single facial parts (i.e., eyes, nose, mouth or contour) only when they are shown within a facial configuration, and a magnitude of priming effects that increase exponentially as the number of face parts presented as primes increases. Patient AR, however, shows both a lack of priming with single natural face parts and a linear priming function with the number of face parts in the prime. Overall, these findings indicate that face recognition normally involves the segmentation of faces into discrete parts, which interact with one another in the construction of global face representations. It is proposed that AR's prosopagnosia is due to a deficit of this integration process.
22 The effects of face inversion on reflexive attention to gaze direction
Chris Kelland Friesen1, Walter F. Bischof1 and Alan Kingstone2
1. University of Alberta
2. University of British Columbia
Previous studies of attentional orienting triggered by nonpredictive gaze direction have typically presented a gazing face as a nonpredictive cue, and then a single sudden onset in the periphery as the target. The standard finding is shorter response time (RT) for a target at the gazed-at location compared to the nongazed-at location at short to intermediate cue-to-target stimulus onset asynchronies, or SOAs (Driver et al., 1999; Friesen & Kingstone, 1998; Langton & Bruce, 1999). In three very different studies, this attentional effect of gaze direction was abolished (Kingstone, Friesen, & Gazzaniga, 2000), reduced (Langton & Bruce, 1999), or delayed (Henderson, Friesen, & Kingstone, 1999) when the gazing face was inverted. These results suggest that face processing interacts with gaze processing. However, in the present study, inverted schematic faces gazing nonpredictively to the left or right produced early and robust orienting, indicating that under some circumstances reflexive attention to gaze direction is not necessarily hampered by face inversion. Possible reasons for these different findings, the relationship between face processing and attention to gaze direction, and likely brain mechanisms will be discussed.
23 Producing spatial descriptions: The effect of functional relations and orientation
Tamsen E Taylor, Christina L Gagné and Roy Eagleson
University of Western Ontario
Before a spatial description can be interpreted unambiguously, the reference frame on which it is based must be determined. This poses a problem, because the viewer-centred and object-centred reference frames share spatial terms (e.g. `right’, `front’). Usually the reference frame is not indicated overtly, however, spatial descriptions are usually understood quickly and effortlessly. How is this possible? There is evidence to suggest that people make inferences based on experience with common objects (default assumptions) when producing and interpreting spatial descriptions. This study extends this idea by examining how background knowledge and visual cues interact. Specifically, if the orientation of objects facilitates the functional relationship between the objects, will this affect reference frame choice? In fact, when describing the spatial relationship between functionally related object pairs (e.g., a teapot and a teacup), participants were more likely to use an object-based reference frame when the spatial relationship facilitated the shared function, while a viewer-based reference frame was preferred when the functional relationship was not facilitated by the spatial relationship
24 How might an artificial neural network represent metric space?
Patricia Boechler and Michael R W Dawson
Biological Computation Project, University of Alberta
One important representational concept in cognitive psychology is a metric space. In this kind of representation, different entities (e.g., concepts) are represented as points in a multidimensional space. Relations among the objects (e.g., similarity between pairs of concepts) are represented in terms of the distance between objects in the space. In a metric space, there are certain constraints placed upon these distance relationships: minimality (the distance from an entity to itself is 0), symmetry (the distance from entity A to entity B is the same as the distance from entity B to entity A), and the triangle inequality (the shortest distance between two entities in the space is a straight line). While the notion of a metric space has a long history in cognition, it is difficult to reconcile with newer, biologically motivated representations, called PDP networks. How might a metric space be represented in such a network? To answer this question, we trained a PDP network of value units to rate the distance between pairs of cities in Alberta. Thirteen different cities were used, and the network was trained to generate a rating from 0 to 10 for each pair. The ratings were constrained in such a way that the network's judgements obeyed the three principles of a metric space. The network required 7 hidden units to learn this task. We analyzed the network to determine its underlying representation. We found evidence for a unique metric representation. Essentially, each hidden unit can be viewed as representing a map of Alberta in 4 dimensional space. The connection weights represented the projections of the 13 cities in the map onto a 1 dimensional axis projected through the 4D space at a particular orientation. The hidden unit responded to distance measures taken along this axis. The combination of responses to 7 different axes allowed the network to determine correct responses to this rating task. These results are discussed in terms of contributing a novel notion of metric space representation. They are also discussed in terms of how this kind of metric representation can be easily manipulated to mediate nonmetric responses (e.g., asymmetric responses of the type often seen in similarity judgement tasks).
25 Effects of the 5-HT1A receptor agonist 8-hydroxy-DPAT on performance on two operant timing schedules
M-Y Ho, T-J Chiang, A S A Al-Ruwaitea, S Mobini, C M Bradshaw and E Szabadi
Psychopharmacology Section, Division of Psychiatry, University of Nottingham
Previous experiments have shown that central 5-hydroxytryptamine (5-HT) depletion disrupts behaviour in some, but not all, timing schedules. We examined the effects of the 5-HT1A receptor agonist 8-hydroxy-2-(di-n-propylamino)tetralin (8-OH-DPAT: 25-200 mcg/kg) on performance on two timing schedules. In Experiment 1, rats were trained to respond on two levers (A and B) in 50-s trials in which reinforcement was provided intermittently for responding on A in the first half, and B in the second half, of the trial (free-operant psychophysical procedure). In Experiment 2, rats were trained to press lever A after a 2-s stimulus and lever B after an 8-s stimulus, and were then tested with stimuli of intermediate durations (interval bisection task). In both experiments, quantitative timing indices were derived from the psychophysical functions (%B responding vs time). In Experiment 1, 8-OH-DPAT displaced the psychophysical curve to the left. In Experiment 2, d-amphetamine increased the Weber fraction without displacing the curve. The results of Experiment 1 are consistent with the proposal that 5-HTergic mechanisms help to regulate pacemaker period. However, the results of Experiment 2 do not support this suggestion. Taken together, the results suggest that different neural mechanisms are involved in timing tasks involving temporal distribution of responding and discrimination of the durations of exteroceptive stimuli.
26 Temporal discrimination and effect of marker duration
Laura Mihaita and Robert Rousseau
University of Laval, Quebec
Temporal discrimination was investigated with filled intervals, empty intervals and a gap within a continuous tone. The effect of marker duration on temporal discrimination was evaluated with empty auditory intervals bounded by markers ranging from 20 ms to 300 ms (E 20 ms, E 100 ms, E 200 ms, E 300 ms). The signals were presented at two base durations (50 ms, 1000 ms) in a single stimulus method. The analysis of the results showed that (1) the empty intervals were better discriminated than filled and gap intervals; (2) the performance decreased with marker duration; (3) the base duration has a important effect on temporal discrimination. This indicated that increasing the marker duration from 20 ms to 300 ms affects temporal discrimination of empty auditory intervals with a base duration of 50 ms and 1000 ms.
27 Visual marking is affected by the attentional blink
Chris Olivers and Glyn Humphreys
University of Birmingham
We can successfully ignore old, irrelevant visual distractors, by inhibiting them before new, relevant information appears. This mechanism has been termed visual marking (Watson & Humphreys, 1997). We present an experiment showing that a secondary task creating a so-called attentional blink (e.g. Raymond, Shapiro, & Arnell, 1992) severely disrupts visual marking. A second and third experiment explore how the attentional blink affects visual marking. We found evidence for a reduction in spatial memory, as well as for reduced inhibitory power. The results support the idea that visual marking is a top- down mechanism, demanding attentional resources. Furthermore, they suggest that the attentional blink disrupts a wide range of processes in the visual system.
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.
Raymond, J. E., Shapiro, K. L., & Arnell K. M.(1992). Temporary suppression of visual processing in an RSVP task: An attentional blink? Journal of Experimental Psychology: Human Perception and Performance, 18, 849-860.
28 Visual search keeps track of where the target is not
E S Olds, W B Cowan and P Jolicoeur
Wilfrid Laurier University
Previously, we interrupted pop-out search before it produced a detection response by adding especially chosen additional distractors to the search display. When pop-out for a colour target is interrupted, it nevertheless provides useful information and assists the processes responsible for difficult search. In another experiment, we moved the target when the extra distractors were added to the display, and the assistance disappeared, suggesting that assistance could be based on information about where the target IS located. New analyses of new RT distributions for target-absent trials indicate that the assistance of difficult search by partial pop-out is based, at least in part, on information about where the target IS NOT located. We relate these results to work on "visual marking" by Watson & Humphreys (Psychological Review, 104, 90-122). We also investigate the effect of set-size and target eccentricity on the assistance.
29 IOR and the repetition of location, identity, and response
Tracy L Taylor-Helmick1 and Michael M P W Donnelly2
1. Dalhousie University
2. Vanderbilt University, USA
Inhibition of return (IOR) refers to slowed responses when a target appears in the same location as a preceding stimulus. In four experiments (E1-E4), we examined the interaction of location repetition with the repetition of non-spatial stimulus dimensions (shape, orientation, response-mapping). We presented two stimuli, S1 and S2, which were horizontally or vertically oriented ovals or rectangles. In E1 and E2, two of the four stimuli mapped onto a speeded left button-press to report the shape (or orientation); the other two mapped onto a right button-press. In E3 and E4, S1 and S2 were presented to one of the four quadrants around fixation. Participants localized S1 and S2 (E3) or just S2 (E4) to the left or right. We observed that IOR is tagged predominantly to a repeated spatial location but that its effects can be modulated by the repetition of other, non-spatial, stimulus dimensions.
30 Exogenous and endogenous selection in partial report
Mark J Fenske and Jennifer A Stolz
University of Waterloo
Previous research (Dixon et al., 1997) has shown that bar-probe partial report performance decreases as the duration of the cue increases with peripheral cues, but not with central cues. Five experiments examined the dependence of this effect on processes of exogenous vs. endogenous attention. Experiments 1-3 demonstrated that cue-duration has the same magnitude of effect for peripheral cues when cuing conditions are presented in either a blocked or a mixed design, or when participants are instructed to treat peripheral cues "endogenously". Experiments 4 and 5 demonstrated that the cue-duration effect is reduced, but not eliminated, when stimulus configurations are changed to encourage or require peripheral cues to be used endogenously. Additionally, Experiment 5 established that the cue-duration effect can be obtained with central cues, challenging the view that the cue-duration effect reflects processes of solely exogenous attention. Results are discussed in terms of conjoint processes of endogenous and exogenous attention.
31 Strategic influences on the time course of exogenous attentional orienting effects
Juan Lupiáñez1, Bruce Milliken2 and M. Jesús Funes1
1. Universidad de Granada, Spain
2. McMaster University
In exogenous attentional orienting tasks, the time to respond to a target following an abrupt onset peripheral cue is measured. Response times to targets following shortly after the cue are typically faster at the cued than at the uncued location, while at longer cue-target intervals the opposite result occurs. Although exogenous cueing effects have been documented in a range of performance tasks, variability in the time course of cueing effects as a function of task has not been studied in detail. We report results from a series of studies which show that the time course of cueing effects does vary depending on task, and that this variability is the product of strategies used to avoid costs that occur when spatial orienting of attention is determined by automatic integrative processes.
32 The influence of non-visual attentional tasks upon the visual motion aftereffect
Robert J Houghton, William . Macken and Dylan M Jones
The motion aftereffect (MAE) is a well-known visual illusion in which, following the continuous observation of drifting contours, subsequently presented static contours appear to drift in the opposite direction. Chaudhuri (1990) showed that if observers are engaged in a separate foveal discrimination task superimposed on a moving textured background, the total duration of the MAE is considerably reduced. This finding was explained in terms of the spatial division of attention.
We have investigated whether suitably demanding non-visual attentional tasks (mental backward counting and auditory vigilance tasks) could also lead to a reduction in the duration of the MAE. Both forms of task were found to reliably decrease the duration of the MAE under specific circumstances; spatial factors were important in determining the level of disruption caused by the auditory vigilance task. These findings can be explained in terms of multimodal representations of attentional objects in space and attentional capacity.
33 Crossmodal attention and the perceptual load hypothesis
Jason Chan, Alex Simm and Charles Spence
Lavie's (1995; JEP:HPP) 'perceptual load' hypothesis of selective attention was tested in both an auditory and crossmodal setting, using an auditory motion after-effect. Left- or rightward moving tones were presented over headphones for 60 seconds, after which participants had to make a 2-alternative-forced-choice response regarding the direction of apparent motion of a static tone. In Experiment 1, participants simultaneously had to discriminate either the case (low attentional-load) or number of syllables (high attentional-load) of a rapid series of visually-presented words. Varying the visual load had no effect on the auditory after-effect, although it was shown to modulate the duration of the visual motion after-effect (Experiment 2; as reported previously by Rees et al., 1997; Science). However, when the words were presented auditorily (Experiment 3), auditory after-effects were shown to be stronger in the low-load condition (judging the loudness of the words) than in the high-load condition (bisyllablic or not). These results support the view that there are separable resources available for the processing of auditory and visual stimuli, at least under certain conditions.
34 Allocation of spare capacity: Opposite effects of search load and perceptual load
Daryl E Wilson, Miya Muroi and Colin M MacLeod
University of Toronto at Scarborough
Lavie and Cox (1997) combined a letter search task with a flanker task to test Lavie's (1995) theory that spare attentional capacity is involuntarily allocated to the processing of irrelevant stimuli. In accord with this hypothesis, they found that increasing task load (i.e., search set size) actually decreased interference from foils. To refine the concept of task load, we differentiated perceptual load from search load. Perceptual load was manipulated by varying the number of letters presented; search load was manipulated by varying the number of search locations, as indicated by precues. Consistent with Lavie and Cox, increases in perceptual load produced decreases in interference. However, increases in search load produced increases in interference. These results suggest that task load is not a unitary construct, and that Lavie's theory regarding the allocation of spare capacity is applicable to manipulations of perceptual load but not to manipulations of search load.
35 Is semantic activation in visual word recognition capacity free?
Matthew Brown, Martha Anne Roberts and Derek Besner
University of Waterloo
Neely and Kahan (1999) assert that, "[semantic activation] from a foveally presented task-irrelevant word should occur full-blown even when attention should be fully allocated to other task-relevant events in the display and this should be so regardless of the type of processing (linguistic or otherwise) that must be performed on those task-relevant events." We report data from two lexical decision experiments that falsify this claim.
Neely, J. H., & Kahan, T. (1999). Is semantic activation automatic? A critical re-evaluation. In Roediger, H.L., Nairne, J.S., Neath, I., & Surprenant, A.M. (Eds.) The nature of remembering: Essays in honor of Robert G. Crowder. Washington, D.C.: American Psychological Association. (in press).
36 The fate of unattended auditory information
Bill Macken and Dylan Jones
The extent to which unattended information is processed has been the focus of considerable debate since the 1950s. In this paper we argue that recent methodological and theoretical devlopments serve to inform this debate. The irrelevant sound paradigm involves exposing participants to auditory inputs while they are performing a demanding short-term recall task. The key findings are that unattended information impinges on the processing of attended information, not on the basis of similarity in content between the two sources, but rather on the basis of similarity in process. Results indicate that basic temporal characteristics relating to order information within the unattended channel are processed and that auditory scene analysis may take place outside the focus of attention. Results are also suggestive that semantic processing takes place within the unattended sound
37 Switch costs without task-switching?
Amelia Hunt, Jason Ivanoff, and Raymond Klein
Dalhousie University, Halifax
A change in task results in detriments to performance. These switch costs have been commonly attributed to the reconfiguration of task set by way of executive control processes. Herein we report the results of an experiment where the task instruction is stable. Responses were based entirely on the non-spatial identity of targets, and spatial location was completely irrelevant to the correct execution of the task. The Simon effect (better performance with spatial correspondence between target and response) was observed. Spatially corresponding and non-corresponding stimulus-response ensembles alternated in a series of eight trials. Although there were no changes in task, we observed robust switch costs at two relatively long response-stimulus intervals. The results call into question the general idea that switch costs must be the result of a reconfiguration of task set.
38 Dual task performance in older adults
Leigh Riby1, Tim Perfect2 and Brian Stollery1
1. University of Bristol
2. University of Plymouth
Previous research has revealed a mixed pattern with regards dual task performance in older adults. Whilst all studies agree that older adults are impaired at dual tasks, only some studies reveal a deficit greater than one would predict from the age-change in single task performance. Kieley (1991) argues that older adults show disproportionate costs of dual tasking when memory is involved. In the present work we examine whether this disproportionate age deficit in dual tasks is found for both semantic and episodic memory. We used a dual task in which participants were required to generate category members or paired associate responses to cues that had just been presented (no load) or to cues that had been presented on previous trials (load). In line with Kieley's expectation there was a disproportionate effect of load for the older adults on the episodic task. A semantic version of this task is underway, and will be presented at the conference.
39 Conscious awareness of stimulus and response in a Posner task.
Helen Johnson and Patrick Haggard
Institute of Cognitive Neuroscience, Department of Psychology, University College London
Conscious awareness of cued location/target location compatibility effects was considered in a Posner (1978) based paradigm. Participants made a speeded simple response to a visual target. They then judged either the time of the stimulus onset or their movement onset, using a methodology devised by Libet (1983). As expected, participants made fastest responses to visual stimuli at a validly cued location. However, judgements of light onset time indicate that the cue/target compatibility effect is not due to any change in conscious stimulus detection. In addition, participants were just as efficient at judging their movement onset across cue conditions. That is, our participants were aware that their responses were delayed on incompatible trials, but did not detect the stimulus any later on incompatible trials.
The locus of the delay is therefore placed between the conscious detection of the stimulus and conscious awareness of response initiation. This is problematic for traditional accounts that place the locus of the compatibility effect at or prior to stimulus detection. A response locus of compatibility effects remains consistent with our data.
40 Effects of a mildly intoxicating dose of alcohol on various components of attention
L A McWilliams, S H Stewart, R M Klein, J R Blackburn and R M McInerney.
University of Manitoba
Previous research suggests that alcohol can have adverse effects on attention, depending upon a host of variables including task characteristics (Streufert & Pogash, 1988). The purpose of this study was to investigate the effects of a mildly intoxicating dose of alcohol on various aspects of attentional function. Forty participants were randomly assigned to one of two beverage conditions: alcohol (vodka + orange juice) or control (orange juice). Following beverage consumption/absorption, participants completed a brief battery of attentional tasks based on well-understood paradigms within cognitive psychology (Blackburn et al., 1999). Beverage condition effects were observed on the Stop Signal, Simon, and Rapid Serial Visual Presentation tasks. The pattern of results suggests that alcohol impairs attentional alertness and possibly attentional filtering, while leaving speed of processing intact. Although no effect of beverage condition was observed on speed of processing, alcohol effects on this parameter of attention might be obtained at higher doses.
41 The effect of set size and task on error patterns in immediate memory performance
Marie Poirier1, Josée Turcotte2, Gerry Tehan3 and Kevin Allen2
1. Bolton Institute
2. University of Laval
3. University of Southern Queensland, Australia
Estes (1991) challenged what he called the "trial-unit model" of short-term recall, according to which short-term memory is reset at the start of each experimental trial. As others, he argued that at the point of recall the number of items competing for output is larger than the number of items presented in the current trial. We investigated these questions within an experiment where set size and task were varied. Participants performed immediate serial recall and immediate reconstruction of order in three set-size conditions. The results show that the pattern of intrusions and protrusions and the effect of set size are different in recall and reconstruction. Results are discussed in terms of the search set that is involved in immediate memory tasks.
42 Do phonological short term memory and episodic memory contribute to performance on complex working memory span tests?
Katy J Lobley, Susan E Gathercole and Alan D Baddeley
University of Bristol
To assess the contribution of phonological short-term memory and episodic memory to complex working memory, three experiments manipulated phonological similarity and level of processing within a listening span test (developed from Daneman and Carpenter, 1980). Phonological similarity was found to impair memory span, regardless of the type of complex span test used (single-word or sentence version) or age group (6 year olds, 8-year olds and adults). In the single word version, the nature of the processing requirement also affected memory. With phonological similar items, span for semantically encoded words was higher than for physically encoded words. The results indicate that complex span tests draw on components of the memory system other than working memory alone, and specifically tap both the phonological loop and episodic memory.
43 Effect of processing temporal or spatial order in short-term memory on concurrent time estimation
Claudette Fortin, Julie Champagne and Marie Poirier
Université Laval, Québec
In this experiment, identification of the temporal or spatial position of a memory probe was performed during a temporal interval production. Memory-set size was manipulated in the temporal and spatial tasks. In a reaction time condition, both tasks produced RT functions with similar slopes, RT increasing markedly with set size. When the memory task was executed during time production however, increasing set size significantly lengthened produced intervals when the memory set was searched for the temporal position of the probe, but not when its spatial position was identified. These results show, with two memory tasks of equivalent difficulty, that the effect from nontemporal processing on concurrent time estimation is dependent on the specific resources involved in the nontemporal tasks. Furthermore, by showing distinct patterns of interference on a concurrent task with identification of temporal and spatial position, they contribute to distinguish memory for temporal order from memory for spatial order.
44 Aging and the ranschburg effect: No evidence of reduced response suppression in old age.
Elizabeth A Maylor1 and Richard N A Henson2
1. University of Warwick
2. Institute of Cognitive Neuroscience, University College London
We report two experiments designed to test one aspect of Hasher and Zacks's (1988) reduced inhibition hypothesis, namely, that old age impairs the ability to suppress information in working memory that is no longer relevant. In Experiment 1, young and older adults were presented with lists of six letters and were asked to recall each list immediately in the correct order. Half of the lists contained repeated items (e.g., BDGDVP) while half were control lists (e.g., VTBGPD). The recall of non-adjacent repeated items was worse than that of control items (repetition inhibition; the Ranschburg effect). There was evidence of a larger Ranschburg effect (i.e., greater response suppression) in older than in young adults. In Experiment 2, young and older adults were required either to recall the list or to report if there was a repeated item in the list. The task for each list (serial recall or repetition detection) was signalled by a post-list cue. Repetition detection performance was high and similar in the two age groups. There was also no age difference in repetition inhibition. When age differences in overall performance were taken into account, however, there was evidence of increased repetition inhibition with age in both experiments. It is concluded that, contrary to Hasher and Zacks's reduced inhibition hypothesis, response suppression involved in serial recall from short-term memory is not reduced by normal aging.
Hasher, L. & Zacks, R. T. (1988). Working memory, comprehension, and aging: A review and a new view. In G. H. Bower (Ed.), The psychology of learning and motivation (Vol. 22, pp. 193-225). New York: Academic Press.
45 Is inhibition induced by the retrieval-practice paradigm?
Josée Turcotte1, Sylvain Gagnon2 and Marie Poirier3
1. Université Laval
2. Université du Québec à Trois-Rivières
3. Bolton Institute
Many studies have illustrated that prior retrieval episodes can make subsequent retrieval of related information more difficult. Recently, Anderson, Bjork and Bjork (1994) proposed a new paradigm where the retrieval process itself induces forgetting. Our aim was to replicate the basic effect of this paradigm with young and older adults. Sixty-four French speaking participants, 32 young and 32 older adults studied a series of category-exemplar pairs (e.g. Fruit-Banana). In a second phase, they practised retrieval of half members of half categories through a category-plus-stem cued-recall test (e.g. Fruit Ba_____). After a retention interval, they executed a final retrieval test where output interference was controlled by a category-plus-stem cued-recall (e.g. Fruit B_____). Results show that the inhibition effect is rather weak in young adults, varies according to the moment of retrieval of 'inhibited category members' and is not observed in older adults.
46 Robust false memory effects under conditions of paired-associate learning
Erin D Sheard and Colin M MacLeod
University of Toronto at Scarborough
Substantial levels of false recall and recognition occur for lists of words semantically related to an unpresented word (Deese, 1959; Roediger & McDermott, 1995). Manipulations that attenuate this "false memory effect" have reduced the probability of the presented words being linked semantically. We employed a paired-associate learning procedure to de-emphasize the relation between items and to make individual items more distinctive. Subjects items either (a) as 12 individual words (standard condition), (b) as 12 words, each paired with a number (number-noun condition), or (c) as 6 two-word pairs (noun-noun condition). If pairing words with numbers minimizes semantic analysis of each item and reduces the probability of forming links between items, the false memory effect should be reduced. However, the results indicated increased false memory in the number-noun condition, relative to the standard and noun-noun conditions, which did not differ. Correlational analyses were used to explore the causes of the false memory effect.
47 Effects of interference in an AB-AC paradigm: Event-related potentials of associative recall
A D Hughes and R Cabeza
University of Alberta
In event-related potential (ERP) studies of episodic retrieval, a parietal positivity has been attributed to recollection, and a frontal positivity, to monitoring. However, in ERP studies of associative recall that used a recognize-then-recall paradigm, the frontal effect was not found and the parietal effect was not attenuated by interference. To investigate these issues, we recorded ERPs during direct cued recall and manipulated interference using an AB-AC paradigm. Participants studied two lists of word-pairs, with half of the pairs in the interference (AB-AC) and half in the control (AB-CD) condition. Following the study phase, subjects recalled either the first or the second list. Preliminary analyses yielded two results. First, the parietal ERP effect was attenuated in the interference condition. Second, ERPs for recalled items, as compared with non-recalled items, were associated with a bilateral frontal positivity. These results are consistent with recollection and monitoring interpretations and extend the ERP literature to include associative recall.
48 Five years later: Children's memory for stressful experiences
Carole Peterson and Nikki Whalen
Memorial University of Newfoundland
Long-term memory for medical emergencies was assessed in children who were between 2 and 13 years of age at the time of injury. The children's injuries included broken bones, lacerations requiring suturing, burns and dog bites-i.e., highly salient, often extremely distressing, unique experiences. Children had been recruited from a hospital Emergency Room at the time of injury and relevant witnesses interviewed shortly afterwards, so the accuracy of the children's recall of both injury and hospital treatment details could be determined. Previous interviews showed that up to 2 years after suffering their injuries, children had forgotten almost nothing of their experiences. The current study tracked down these same children three years after their last interview (and 5 years after their injury), and re-interviewed them about what they recalled. Although most research involving such long-term memory has found children to recall relatively little, the children in this study had extremely good recall in spite of the passage of 5 years between original injury and final recall. Implications for children's eyewitness testimony are discussed.
49 Aging and goal pursuit: The effects of implementation intentions on prospective memory in older adults.
Alison L Chasteen1, Denise C Park2 and Norbert Schwarz2
1. University of Toronto
2. University of Michigan, USA
Recent work on goal completion has found that people achieve their goals more successfully when they form implementation intentions, which are intentions associated with specific situational cues (Gollwitzer, 1999). Gollwitzer (1999) has argued that implementation intentions rely on automatic processes, which suggests that older adults might benefit from forming implementation intentions because automatic processes in memory are relatively age-invariant (Jacoby, Jennings, & Hay, 1996). The purpose of the present study was to determine whether forming implementation intentions would help older adults with tasks requiring event-based prospective memory. All of the older participants completed two types of prospective memory tasks. One task was a background pattern (BP) task in which they had to remember to press a key whenever a specific background pattern appeared on the computer screen. The other was a day of the week (DOW) task in which they had to remember to write the day of the week in the upper right corner of any sheet of paper they received. Older participants were randomly assigned to one of three conditions: (1) form an implementation intention for the BP task, (2) form an implementation intention for the DOW task, or (3) form no intentions for either task (control). No effects were found for the BP task, but an implementation intention effect was found for the DOW task. Participants who formed implementation intentions for the DOW task remembered to write the day of the week much more often than participants who did not form intentions for the DOW task. These results suggest that forming implementation intentions does improve older adults’ prospective memory for some tasks.
50 Retrospective revaluation in humans.
M E Le Pelley and I P L McLaren
University of Cambridge
The phenomenon of retrospective revaluation has been a challenge to many associative learning theories as it involves a change in the associative strength of a cue on trials on which that cue is absent. Most current theories of retrospective revaluation explain it in terms of new learning about the absent cue on trials in which that cue is absent but expected (e.g. Dickinson & Burke¹s (1996) modified SOP model; van Hamme & Wasserman¹s (1994) modified Rescorla-Wagner rule). We instead propose a more memory-based explanation of retrospective revaluation, which we suggest results from changes in the retrievability of previously learnt information as a result of subsequent experience. The present series of experiments attempts to distinguish between learning-based and memory-based theories of retrospective revaluation, focussing on Dickinson & Burke¹s (1996) modified SOP model, and a version of McLaren¹s (1993) APECS network.
Dickinson, A., & Burke, J. (1996). Within-compound associations mediate the retrospective revaluation of causality judgements. Quarterly Journal of Experimental Psychology, 49B, 60-80.
McLaren, I. P. L. (1993). APECS: A solution to the sequential learning problem. Proceedings of the XVth Annual Convention of the Cognitive Science Society.
van Hamme, L. J., & Wasserman, E. A. (1994). Cue competition in causality judgments: The role of nonrepresentation of compound stimulus elements. Learning and Motivation, 25, 127-151.
51 Mechanisms of predictive and diagnostic causal inferences
F J Lopez1, P L Cobos1, A Caño1 and David Shanks2
1. University of Malaga, Spain
2. University College London
In predictive causal inference people reason from causes to effects while in diagnostic inference they reason from effects to causes. Independently of the causal structure of the events, the temporal structure of the information provided to a reasoner may vary, e.g., multiple events followed by a single event versus a single event followed by multiple events. Here we report 2 experiments in which the causal structure and the temporal information were varied independently. The results reveal that inferences were influenced by the temporal but not by the causal structure of the task. The results are relevant to the evaluation of two current accounts of causal induction, the Rescorla-Wagner (Rescorla & Wagner, 1972) and Causal Model theories (e.g., Waldmann & Holyoak, 1992).
52 A developmental study of contingency perception and its relation to negative affect
Margo C. Watt1 and Peter J. McLeod2
1. Dalhousie University
2. Acadia University
The present study investigated age and gender differences in contingency perception and its relation to negative mood. The study combined a mixed prospective and cross-sectional design that allowed assessment of students' affective responses to both experimentally-induced and naturally-occurring stressors. A total of 335 students completed a Judgement-of-Control Task in which the contingency parameters of sufficiency and necessity had been manipulated, an Induced Failure Task, and measures of negative affect and stress. Results provided support for the view that illusory contingency declines with age. Sensitivity to response probability (sufficiency) and random outcomes (necessity) increased with age. Males displayed greater sensitivity to the probability of random outcomes than females. Results indicated that illusory contingency confers some protection from the negative effect of stress but that its role differs between males and females. No support was found for either a moderational or mediational role for contingency perception in the development of negative affect.
53 Human sequence learning: Evidence for an underlying associative and/or cognitive process
University of Cambridge
A three-choice serial reaction time task (SRT) was used to investigate human sequence learning. After the task, interviews were carried out to see whether subjects have become cognitively aware of the underlying rules. There is evidence that people's reaction times become significantly faster on rule-consistent stimuli when compared to rule-inconsistent ones. This also holds when people generalise to novel sequences. However, the interviews generally showed that most people were only able to verbalise a part of the underlying rule. The simple recurrent network (SRN) as introduced by Elman (1990) provides the opportunity to model sequence learning in a simple associative fashion. Therefore, the SRN was trained on the same sequences as our human subjects. While the SRN could learn the sequences, it was never able to generalise to the rule-consistent novel sequences. It needs to be said that this is not due to a particular deficit of the SRN. Simulation experiments in our laboratory have shown that, all other factors being equal, the SRN is able to generalise to rule consistent sequences in other sequential tasks. In our case, we have a task that can entirely be solved by humans, but not by the SRN. As a result, we may need to think about a more complex associative process. At the moment, other experiments are carried out in order to shed light on the learning process in question. The results and conclusions of those experiments will also be discussed at the conference.
Elman, J.L. (1990). Finding structure in time. Cognitive Science, 14, 179-211.
54 A qualitative dissociation in sequence learning.
F W Jones and I P L McLaren
University of Cambridge
There is controversy as to whether implicit learning phenomena are due to the operation of a dissociable unconscious learning process. This is largely as a result of the lack of consensus over how we should measure awareness. However, recent research has tended to concentrate less on the question of awareness and more on other characteristics of implicit learning (cf Buchner & Frensch, 1997). The experiments reported here continue in this vain. To be more precise, we sought to determine whether the same sequences could be learnt in a qualitative different manner if they were presented under different training conditions; a result that might provide some support for the existence of dissociable learning processes. The experiments employed a two choice RT task in which the order of stimulus presentation was determined by concatenating the following sequences: RRR, LLR, LRL, RLL (R=right, L=left). If the subjects were asked to look for a pattern and given a marker to indicate the beginning of each triplet of trials, then they typically found RRR the most salient sequence and the easiest to learn (as indexed by the greatest RT difference in comparison to a group that controlled for sequential effects). However, if they were not informed about the sequences and no markers were provided, then a different pattern of learning resulted, with RRR being learnt least. Possible explanations of this dissociation are that: (1) it reflects the operation of two different learning mechanisms; (2) it arises because subjects chunk the sequence differently in the two cases. The relative merits of these possibilities are discussed.
Buchner, A., & Frensch, P.A. (1997). Sequence learning: Phenomena and models. Psychological Research, 60, 1-3.
55 Category creation - association or competition?
Mark Suret and I P L McLaren
University of Cambridge
In the past, much time and effort has been devoted to the study of category learning in humans, and this in turn has generated models of categorisation which have tended to be associative in nature. The question of category formation as opposed to category acquisition has received relatively little attention, and this has been the main focus of this research, based on the work of Wills and McLaren (1998). We now know that people are able to extract category structure without explicit instruction, and we have a paradigm (free classification) for studying this process. The differences between free classification and categorisation-with-feedback have been investigated by training or pre-exposing to one set of stimuli and then testing generalisation to novel stimuli drawn from that set.
Results so far show that, remarkably, free classification may, under some circumstances, be a better method of training. Although response performance on test does not vary with training, the reaction time data indicates that using a free classification paradigm may decrease the time taken to reach a decision on the generalisation test. The inverted U-shape of the reaction time curve for generalisation, displayed by groups trained with feedback, is greatly reduced in curvature and absolute response time for the free classification group.
The results generated from the experiments have been compared with two different neural network architectures, error correcting and competitive, to discover how category learning in humans might proceed.
Wills, A.J., and McLaren, I.P.L. 1998. Perceptual learning and categorisation: From ignorance to consistent classification without explicit instruction. Quarterly Journal of Experimental Psychology, 51B
56 What's in a fingerprint? A PCA approach to fingerprint identification and categorisation
Jason M Tangen, John R Vokey and Lorraine G Allan
Eigenvectors of an autoassociative memory matrix, derived via singular value decomposition of a set of pixel-maps of fingerprints, contain information that is useful for predicting various traits of the individuals from whom the patterns were obtained. When the projections of the fingerprints onto the eigenvectors were used as input, the model was able to identify each of the patterns it was trained with, as well as generalising its learning to novel prints not presented in the training set. Eigenvectors are ordered by the amount of variance they explain, and reflect the visually based categorical structure of the fingerprint patterns they represent. Thus, it is the individual patterns as a whole and not the specific aspects of their content that are discriminated into discrete categories. The results are discussed in terms of the advantages of this approach to characterising both perceptual/conceptual learning and the stimuli used to investigate it.
57 The effect of phonotactics and neighbourhood density on spoken word recognition in preschoolers
John Logan and David Ridgeway
Vitevitch and Luce (1999) provided evidence that phonotactics and neighbourhood density make independent contributions to spoken word recognition in adults. This finding has implications for spoken word recognition in children. The limited size of children's lexicons suggests that neighbourhood effects may be more limited and phonotactic effects more prevalent in children as compared to adults. In the present investigation we evaluated the contribution of neighbourhood density and phonotactics in a word recognition task. Stimuli consisted of 56 spoken words varying in neighbourhood density and phonotactic probability selected from a child-based lexicon. 2-, 3-, and 4-year-old children were asked to identify the spoken word by pointing to the word's referent in a 4-picture display. The 2-year-old's recognized words with high probability phoneme sequences more accurately than low probability sequences. In contrast, the effects of neighbourhood density begin to manifest themselves later during childhood as additional words are added to the lexicon.
58 Pseudohomophone base-word frequency and lexicality effects
William J Owen and Ron Borowsky
University of Saskatchewan
The effect of base-word frequency on pseudohomophone (PH) naming latency has produced equivocal results. For example, the studies of McCann and Besner (1987), Marmurek and Kwantes (1996), and Herdman, LeFevre, and Greenham (1996) reported no frequency effect, a negative frequency effect, and a positive frequency effect, respectively. However, different ratios of lexical to non-lexical stimuli were used in these studies, and the validity of the PH stimuli was not examined. A new set of PHs, validated separately for each participant, was tested in pure and mixed stimulus blocks. A PH advantage over nonword naming latency was found only in the mixed block condition, and a frequency effect was observed only in the pure block condition. Comparisons were also made to other PH stimuli sets. The importance of list structure, naming instructions, and PH validation are discussed along with implications of PH naming for models of lexical access.
59 Basic processes in reading: Semantics affects speeded naming of high frequency words in an alphabetic script
Bahman Baluch1 and Derek Besner2
1. Middlesex University
2. University of Waterloo
Previous work on single word naming in university level readers has shown that semantics affect the naming of low frequency words in both an alphabetic script like English, and a syllabic script in which the spelling-sound correspondences are consistent (Japanese Kana). The present experiment shows that a semantic factor (imageability) affects naming time to both low and high frequency words in an alphabetic script (Persian) when the word is opaque but not when it is transparent. The characteristics of opaque words that promote the use of semantics are discussed. At least in some orthographies, semantics play a larger role in single word naming
than previously thought.
60 When parallel processing in visual word recognition is not enough: new evidence from naming
Martha Anne Roberts, Kathleen Rastle, Max Coltheart and Derek BesnerUniversity of Waterloo
Low frequency exception words are named more slowly and are more error prone than are low frequency regular words (the regularity effect). Rastle and Coltheart (1999) reported that this irregularity disadvantage is modulated by the position in the word at which the irregularity occurs such that words with early irregularities are named more slowly than words with late irregularities. They argued that these data implicate serial processing, and provided a successful simulation with the DRC model of reading aloud (a model which has a serial component). However, Zorzi (in press) also simulated these data with a model that operates solely in parallel (Zorzi, Houghton, & Butterworth, 1998). Thus it is not clear that Rastle and Coltheart's (1999) findings implicate serial processing. New data are presented that adjudicate between these models. An interaction between regularity and serial position of irregularity is reported for human readers. This interaction is simulated by DRC but not by the Zorzi et al., (1998) model. These results are discussed in terms of the different properties of the stimuli used in Rastle and Coltheart (1999) versus those used in the current study.
Rastle, K. & Coltheart, M. (1999). Serial and Strategic Effects in Reading Aloud. Journal of Experimental Psychology: Human Perception and Performance, 25, 482-503.
61 Prime validity modulates masked priming
Glen E Bodner and Michael E J Masson
University of Victoria
If a masked prime automatically preactivates its lexical representation, then masked priming should not be affected by the proportion of different types of primes in the stimulus list. Contrary to this claim, we report that masked repetition priming in the lexical decision task with 45-ms and 60-ms primes is typically greater when the proportion of repetition prime trials is .8 rather than .2. Moreover, contrary to the notion that consciously-deployed strategies such as expectancy drive proportion effects observed with clearly visible associative primes, we found that masked associative priming with 45-ms primes occurred only when a set of filler trials had repetition rather than unrelated primes. We suggest that masked priming reflects the retrieval of the prime's episodic memory trace, and that this retrieval is modulated in a non-conscious way by the list-wide validity of the priming stimuli.
62 Negative priming for homophone and pseudohomophone pairs: Further support for a retrieval-based account of negative priming.
Penny A MacDonald1, Colin M MacLeod1 and Ken N Seergobin2
1. University of Toronto at Scarborough
2. McMaster University
We tested negative priming for homophones and pseudohomophones. In Experiment 1, we compared negative priming for homophones in a standard negative priming task, where distractors are encoded shallowly, with negative priming in a modified task, where distractors are encoded deeply. We observed negative priming for homophones in the shallow encoding task only. In Experiments 2 and 3 negative priming occurred for lexically related word pairs and lexically similar word-pseudohomophone pairs but not for lexically dissimilar word-pseudohomophone pairs. Negative priming arises for homophones and pseudohomophones in the shallow encoding task because the perceptual similarity between the prime distractor and probe target causes a reinterpretation of the prime item (Kahneman, Treisman, & Gibb, 1992; Loftus & Loftus, 1974) and not because the associated responses are the same. These findings support a retrieval-based explanation of negative priming, placing greater emphasis on probe processing than traditionally has been ascribed.
63 Do "bilingual" infants use fine phonetic detail in word learning tasks?
Chris Fennell and Janet Werker
University of British Columbia
Fourteen-month-old monolinguals confuse phonetically similar words in a word-object association task (Stager & Werker, 1997); however, older, more proficient word-learning infants do not (in prep). It has been hypothesised that 14-month-olds fail because linking words to objects is difficult at the beginning stages of word learning, leaving infants with insufficient attentional resources to listen closely to the words. Extending this hypothesis to bilinguals raises two possibilities: elimination of the listening strategy because more phonetic detail is needed to discriminate words in two languages, or retention of the strategy because of the increased cognitive load of learning two languages. Bilinguals of 14 months were tested in the word-object association task. Bilingual exposure was assessed with a structured parental interview. Analogous to the monolinguals, the 14-month bilinguals confuse similar sounding words. Research with 17- and 20-month bilinguals is now underway to clarify if the monolingual pattern remains consistent or diverges with development.
64 Preschoolers' and adults' interpretations of familiar and novel adjectives
Adam McCrimmon, Andrea N Welder and Susan A Graham
University of Calgary
In Experiment 1, we examined preschoolers’ and adults’ understanding of the referential scope of familiar adjectives. On each trial, we labeled a target animal with one of three different types of adjectives: a transient emotional state term, a transient physiological state term, or a stable trait term. Participants were asked whether these terms could apply to a subordinate match, a basic match, a superordinate match, or an inanimate object. Results indicated that 4-year-olds and adults extended the stable trait terms, but not the emotional or physiological terms, to the subordinate and basic matches. In Experiment 2, we presented 4-year-olds and adults with novel adjectives in one of two syntactic frames ("This X is daxy" vs. "This X feels daxy"). Participants were more likely to extend the novel adjective to subordinate matches, if they were in the "is" condition than if they were in the "feels" condition. We will discuss these findings in terms of implications for children's acquisition of differing types of adjectives.
65 Poor comprehenders can integrate and retain semantic information from heard text
Alice Spooner, Sue Gathercole and Alan Baddeley
University of Bristol
Three experiments compared seven- and eight-year-old poor comprehenders with normal readers and poor decoders on an immediate and delayed sentence recognition task. Internal reference to integrative semantic representations was indexed by the inflated recognition errors of foil sentences that were semantically congruent with original story information . Poor comprehenders did not differ from skilled comprehenders either in the degree to which they integrated information, as estimated by immediate recognition of single-story information. Furthermore, the groups were not differentially affected by memory demand in their retention of story information. Possible explanations for these findings include underestimation of comprehension ability in selection, and the short simple materials of these experiments being within the ability of poor comprehenders.
Daneman, M., & Carpenter, P. A. (1980). Individual Differences in Working Memory and Reading Comprehension. Journal of Verbal Learning and Verbal Behavior, 19, 450-466.
67 Case studies in dyslexia
Catherine G Penney
Memorial University, St. John's
Individual case studies are reported of dyslexic students aged 9 to 16 whose reading achievement has been monitored for up to five years. Increases in reading achievement are compared for control periods, during which students received no special help outside of school, and for experimental periods during which they recived invididualized tutoring using a rime-based method of teaching word identification (Glass Analysis). The glass analysis word identification and spelling drills consist of rehearsing the associations between orthographic patterns and pronunciation for rimes or complete syllables embedded in words. Students also did oral reading of text passages containing the words that had been drilled. Reading achievement, as measured by standardized tests, improved more during experimental periods than during control periods, and some benefit was obtained for spelling as well. Although comprehension was not explicitly taught, reading comprehension frequently improved as word identification improved.
68 Dissociating performance in written and oral calculation: Evidence from right hemisphere developmental learning disability
Sandra Martin-Chang1 and Annalena Venneri2
1. McMaster University
2. University of Aberdeen
Spatial acalculia, the inability to properly arrange numbers during computation but preserved calculation principles, is reported frequently following right brain damage. We report on the case of a 9.8 year old child, C.G., who was diagnosed with right hemisphere developmental learning disability prior to testing. He demonstrated his intact arithmetic skills by correctly answering questions concerning number comprehension, number facts, and sign recognition. He could also correctly process verbal addition and subtraction questions. However, C.G. showed no appreciation for the spatial arrangements of numbers. The striking dichotomy of his performance was that he could produce the correct answer to difficult arithmetic questions, while he was unable to perform simple written questions. This case study supports Hecaen et al.'s (1961) claim that the right hemisphere contributes spatial knowledge to arithmetic. Developmental defects in spatial processing can result in spatial acalculia, while allowing normal acquisition of other aspects of calculation.
69 Electrophysiology of emotional memory: Evidence for the valence hypothesis
F Dolcos and R Cabeza.
University of Alberta
We investigated two event-related potentials (ERPs) phenomena: (1) ERPs tend to be more positive for emotional pictures than for neutral pictures (emotion effect); (2) ERPs to be more positive for subsequently remembered than for subsequently forgotten items (Dm effect). We recorded ERPs while subjects rated the pleasantness of pleasant, unpleasant, and neutral pictures, and later, we tested their recall for information in the pictures. Preliminary analyses yielded two main results. First, the emotion effect on frontal electrodes was left-lateralized for pleasant pictures but right-lateralized for unpleasant pictures. This dissociation is consistent with the valence hypothesis of hemispheric lateralization of emotion. Second, the Dm effect occurred earlier for emotional (400-600 ms) than for neutral pictures (after 600 ms). This earlier Dm effect could account for recall performance, which was better for emotional than for neutral pictures. The results established a link between two seemingly unrelated ERP phenomena: the emotion effect and the Dm effect.
70 Priming the meaning of homographs in children with autism
Suzanne Hala, Penny Pexman and Christa Leibel
University of Calgary
Several theories have arisen to account for the core deficit underlying autism. The research reported here was designed to provide a specific test of two of these theories Weak Central Coherence theory and Executive Dysfunction theory. In this research we investigated how meaning is accessed in a reading task in both normally developing children and children with autism. In support of the Weak Central coherence view, previous research has found that children with autism fail to take account of context to guide pronunciation of homographs embedded in sentences (Happé, 1997). The embedded sentences task, however, also makes significant executive demands on the reader. In this research we sought to reduce executive demands by testing the use of meaning in a priming task. Participants were first presented with a word prime that was associated with either the dominant or subordinate pronunciation of the subsequent target homograph (e.g. RIP-TEAR). Participants were asked to pronounce both words and responses were scored for correct pronunciation and reaction time. The results will be discussed in relation to the two competing theories.
71 Multiple neuroanatomical features of high functioning autism: A structural MRI study.
Patricia Cowell1, Matthew Howard2, Jill Boucher3, Neil Roberts2, Andrew Mayes1 and Paul Broks1
1. Department of Human Communication Sciences, University of Sheffield
2. Magnetic Resonance and Image Analysis Research Centre, University of Liverpool
3. University of Warwick
High resolution MRI scans were acquired for ten adult males with autism and ten healthy control subjects matched for age, sex and verbal IQ. Unbiased estimates were obtained for right and left hemispheric volumes of the medial temporal lobe (MTL: amygdala, hippocampus, parahippocampal gyrus1) and subfields of the prefrontal cortex (PFC: dorsal and orbital) using the Cavalieri method of modern design stereology. Intra-cranial volume (ICV) was also measured. Multivariate anovas were performed separately for MTL and PFC using Diagnosis (Autism/Control) as a grouping factor and Region as a repeated measure. In MTL, Diagnosis interacted with Region (F=7.52, d.f. 2,34, p<.01). Post-hoc t-tests showed this was due to larger amygdala volumes (p<.05) and smaller hippocampal/parahippocampal volumes (p’s<.10) in patients compared to controls2. In PFC, Diagnosis interacted with Hemisphere (F=4.68, d.f. 1,18, p<.05) and Hemisphere x Region (F=3.62, d.f. 1,18, p<.10). Right dorsal PFC was larger than the left in controls (p<.05) but not patients with autism. ICV was larger in patients than controls (F=4.48, d.f. 1,18, p<.05), but sporadic correlations with regional brain measures did not support systematic use of this factor as a covariate. Results provided support for a neuroanatomic profile in patients characterised by: (a) autism-specific effects in the MTL; and (b) PFC asymmetry similar to that observed in other neurodevelopmental disorders.
(1) Note: one patient with autism had missing parahippocampal data and was deleted from all MTL analyses. (2) Howard, M.A., Cowell, P.E., Boucher, J., Broks, P., Mayes, A. and Roberts, N.A structural magnetic resonance imaging (MRI) study of adults with Asperger’s syndrome. Proceedings of the Annual Human Brain Mapping Meeting, submitted.
72 Stability of frontal electroencephalogram (EEG) measures and cognitive and affective development in children
Carrie E Sniderman and Louis A Schmidt
Although the pattern of resting ongoing frontal brain electrical activity (EEG) has been consistently linked to individual differences in cognitive and affective processes in infants, children, and adults, few studies have examined the temporal stability of electrocortical measures in pediatric populations in relation to developmental outcome. In the present study, we examined 1) the stability of regional EEG power and asymmetry measures in a group of 6 year-old children prior to and upon entry into grade school, and 2) the relations of these measures to cognitive and affective development. We found excellent test-retest stability in regional EEG power measures and high stability in regional EEG asymmetry measures across time in frontal EEG power and asymmetry measures. We are currently examining children whose pattern of resting frontal brain electrical activity remains stable in predicting their cognitive abilities and socio-emotional development during the early school age years. Findings are discussed in terms of the utility of regional EEG measures, which are relatively noninvasive, in understanding individual differences in cognitive and emotional development.
73 Differences in reported test anxiety of native vs. non-native english speakers
Antonia Mantonakis and Douglas A Bors
University of Toronto at Scarborough
Using 966 university students, the present study examined the effects of gender and first-language (as an indication of culture) on test anxiety using the Spielberger Test Anxiety Inventory (TAI). There was a main effect for gender (women reported more test anxiety than men did), a main effect for first-language (non-native English speakers reported more test anxiety than native English speakers did), with no interaction between gender and first-language. Neither gender nor first-language, however, affected the factor structure of the TAI. Confirmatory factor analyses indicated that a two-factor correlated model, comprising cognitive aspects of anxiety (worry) and bodily arousal (emotionality), best fit all groups. The two factors, nonetheless, were highly correlated (.80), arguing for a one-factor model. Examining participants' correlations between worry and emotionality scores and results in Introductory Psychology and Raven's Advanced Progressive Matrices failed to demonstrate a consistent pattern from which to choose between the one-factor and two-factor models.
74 Temporolimbic functions: Anxiety, depression and cognitive flexibility in university students
Philip A Murphy Jeremy M Barry and Peter G Henke
St. Francis Xavier University, Antigonish
University students completed the Limbic System Checklist (LSCL-33) prior to an assessment of anxiety measures. The upper quartile LSCL-33 participants had higher scores on both trait and state anxiety. Participants who were first administered cognitive/behavioral tasks exhibited greater startle responses. In a second study, the upper quartile LSCL-33 participants had higher scores on the Beck Depression Inventory (BDI-2), and they also used indirect solutions more frequently in solving the so-called Luchins Water Jar Problem. The findings support the hypothesis that LSCL-33 scores are related to measures of anxiety, depression, and cognitive flexibility.
75 Raven's advanced progressive matrices test is a measure of 'G': Evidence in support of a single factor model
Tonya L Stokes and Douglas A Bors
University of Toronto at Scarborough
The Raven's Advanced Progressive Matrices test (APM) was originally designed as a nonverbal measure of Spearman's g (general intelligence). Much disagreement exists, however, over the cognitive components responsible for performance on this task. While Spearman believed g represented a single factor, many have argued that the items of the APM are multidimensional in nature. Recently, DeShon et al. (1997) argued that most of the items can be classified as either analytical or spatial in nature. The purpose of the current study was to further examine the dimensionality of the APM and to test DeShon et al.'s factorial structure. Five hundred and six first-year university students completed the APM using standard timing and instructions. A confirmatory factor analysis revealed that a one-factor solution was a better fit than the orthogonal two-factor solution. Furthermore, the two-factor correlated model was no better of a fit than the one-factor model.
76 Motor priming and the cross modality problem
Alick Elithorn, David Jones and Mary Norrish
Children's Hope Foundation, Birkbeck, PRIME.
In 1962, George Ettlinger and later Noble, carried out experiments with rhesus monkeys designed to elucidate that an intolerance of ambiguity plays a role in determining natural dyslexia. Subsequently Ettlinger went on, in a series of experiments with both animals and humans, to delineate some of the difficulties that a bilateral but hierarchical system has in intersystem communication.
Recently, with the aid of a domestic computer, we have carried out a number of experiments which simulate in man some of the procedures developed by Ettlinger. These experiments, which examine an individual's preferences within a range of mirror figure matching procedures, are described and some results presented in detail, together with alternative statistical analyses.
In experiments undertaken with adults and children, the results are interpreted in relation to the development of reading skills. In three adult subjects, the procedures have been used to reveal some drug-sensitive changes in cross-modal communication (confusion ?) which appear to determine periods of intense mental dysfunction.
We relate these results to a biological basis for personality differences described by Sheldon and Galton and to individually variable procedures which subjects adopt in seeking to match patterns (letters). These individual differences in behaviour are related to theories of brain structure developed by Galton and Sheldon. and to more detailed modern theories involving such conepts as those underlying "feature analysis" and the rotation of "internal images".
Sheldon in his consensual tripartite physiology of personality describes evolutionary infrastructures which could underpin Galton's more psychological concept of sensory types. Galton's thinking may be perhaps, in turn, reinterpreted as representing phenotypes as genetic clusters within the human genome.
At a more tentative level, it is suggested that the body of Ettlinger's work, together with the present results, define intersystem relationships which cannot be transformed into the algebra of language. The results are therefore held to support the theories of Herder and other members of the Enlightenment- eg Boule that language is an algebra which damages rather than supports pure thought.