Brain and Cognition 54 (2004) 128132 www.elsevier.

com/locate/b&c

Abstracts

The Rotman Research Institute 13th Annual Conference

March 1718, 2003, Toronto, Canada Part 1: Abstracts of presentations

1. The importance of behavioural data in interpreting neuroimaging activation maps

Ravi S. Menon Canada Research Chair in Functional Neuroimaging, Robarts Research Institute; Professor of Medical Biophysics, Diagnostic Radiology, Physics and Psychiatry, University of Western Ontario, Ont., Canada Functional MRI has been very useful in elucidating the neural substrates for dozens of cognitive tasks. Pretty coloured maps of statistical probabilities can be found in many journals. In isolation, these maps are only of limited utility. In our lab, we use extensive intrumentation to collect behavioural data in the magnet, so that we can relate human performance with the magnitude and timecourse of the fMRI signal. Such an approach allows us to begin to understand how the system as a whole (the brain) goes about processing an input and producing a motor output. Examples from contrast sensitivity detection and visuo-motor learning will be presented.

brain area in which the mirror neurons were rst discovered. Somatosensory cortices seem to be closely connected to the MNS, possibly contributing to dierentiation between brain activations associated with the subjects own movements and those elicited by observation of other persons actions. Our MEG recordings suggest that the human MNS is activated in a well-dened temporal sequence within about 250 ms: visual cortices STS inferior parietal cortex Brocas area motor cortex. Brocas area is activated signicantly stronger during on-line imitation than during execution or observation of nger and mouth movements, emphasizing its role in imitation. The characterization of the human MNS arises some intriguing speculations about the evolutionary and functional relationship between human speech and orofacial/hand gestures. Future studies should address the generality of the MNS in humans and also determine to which extent the MNS functions can be trained. Moreover, it would be important to evaluate whether impaired mind reading and imitation skills, observed in some neurological and psychiatric diseases, could be related to dysfunctional MNS.

2. Towards studies of the social brain: The human mirror-neuron system

Riitta Hari Professor of the Academy of Finland; Head of the Interdisciplinary Brain Research Unit, Low Temperature Laboratory, Helsinki University of Technology, Helsinki, Finland Humans are the most important stimuli for the other humans. Because of the high complexity of such stimuli, direct studies of the eects of other persons acts on the observers brain functions have seemed, for a long time, to be beyond strict scientic scrutiny. The goal of this presentation is to illustrate that some aspects of brain basis of social cognition can be studied with modern brain imaging tools. We copy other persons actions during our whole life, most of the time eortlessly and unconsciously. We also continuously read intentions of others on the basis of their motor acts, postures, and gaze and use that information for predicting their behavior in social interactions. Recent data suggest that this type of automatic reading of other minds could be supported by mirror neurons, rst identied and characterized in the monkey brain. The mirror-neuron system (MNS) seems to match action observation and execution and thereby play an important role in understanding the meaning of actions made by others. Thereby, the MNS could form a link between the sender and receiver of a motor act-based message. We have demonstrated with whole-scalp magnetoencephalographic (MEG) recordings that the human brain contains a MNS comprising a network of several brain areas. These regions include, e.g., the primary motor cortices, the superior temporal sulcus (STS) regions, and Brocas area, the human counterpart of the monkey 0278-2626/$ - see front matter 2003 Published by Elsevier Inc. doi:10.1016/S0278-2626(03)00267-7

3. Exploring sensorimotor interactions by imaging and stimulating the human brain

Tomas Paus Associate Professor, Department of Neurology and Neurosurgery, Department of Psychology, McGill University, Montreal, Que., Canada; Cognitive Neuroscience Unit, Montreal Neurological Instititute, Montreal, Que., Canada In this lecture, I will review three areas of our current research aimed at elucidating the neural mechanisms of sensori-motor interactions. The rst topic concerns the inuence of complex visual and auditory input on motor excitability; this work relates to the motor-resonance theory of Rizzolatti and colleagues. I will describe two experiments, carried out with Dr. Kate Watkins, in which we examined the eects of seen and heard speech on the excitability of the face representation in the primary motor cortex and the role of the Brocas area in mediating these eects. The second topic addresses the question of visual functions of the frontal-eye eld (FEF). I will discuss the results of two experiments, carried out with Dr. Marie-Helene Grosbras, in which we demonstrated the importance of the FEF for visual detection and awareness. The third topic concerns the role of the superior parietal cortex in the awareness of self-generated movements. I will describe an experiment carried out with Dr. Penny MacDonald. Taken together, these experiments suggest that there are no boundaries between action and perception.

Rotman Research Institute Abstracts / Brain and Cognition 54 (2004) 128132

129

4. Processing of music and speech by the human auditory cortex: Neuroimaging evidence
Robert J. Zatorre Cognitive Neuroscientist, Montreal Neurological Institute, McGill University, Montreal, Que., Canada The human auditory cortex has evolved specialized mechanisms for processing the information contained in our acoustic environment. Among the more complex and interesting of these signals are speech sounds and musical patterns. I shall present data bearing on the organization of the human auditory cortices as they pertain to the processing of these types of sounds using functional neuroimaging techniques. In particular, I will argue that studying musical processes oers useful insights into the functional dierences between auditory cortical regions within and between the cerebral hemispheres and how these dierences may relate to higher-order processing of music and speech. Second, I will present data pertaining to anatomical organization, which may help explain hemispheric functional dierences. In particular, I will discuss recent data indicating that individual dierences in white-matter distribution predict speed of learning of novel speech sounds. Together, the data contribute towards a better model of how auditory cortical organization underlies auditory cognitive processes.

proprioception and can be used in order to rene existing skills. The recognized discrepancy pathway relates to a specic perceptual quality and cues in one modality can help to process information in another one. Furthermore, the information from dierent modalities can interact and add new quality of perception that conveys information not inherent in each single modality. In general, it seems that the perception of cross-modal cues, involving cross-modal plasticity mechanisms, is important for playing an instrument and primary cortices that have been classically thought to be unimodal might actually by multifold.

6. Neural concomitants of auditory scene analysis

Claude Alain Scientist, The Rotman Research Institute, Baycrest Centre for Geriatric Care, Ont., Canada; Associate Professor, Psychology, University of Toronto, Ont., Canada Our typical environment is acoustically complex. At any given moment, we might be surrounded by a number of sound-generating elements such as a computer fan, a radio playing music, or a group of people speaking, and several of these elements might be operating simultaneously. In order to make sense of this environment, we must identify, group, and segregate these elements into separate mental representations called auditory streams or objects. This process is known as auditory scene analysis and involves perceptually organizing our environment along two axes: horizontal (time) and vertical (frequency). Organization along the horizontal axis entails sequential grouping of acoustic data over time, whereas along the vertical axis, acoustic elements from simultaneous sound sources are grouped and segregated based on frequency patterns. This presentation will include a brief overview of the acoustic cues that are important for auditory scene analysis. We will outline the theoretical framework that has guided our research over the past several years and will present a number of experiments that have tested and rened this framework. Specic ndings concerning the sequence of cognitive and neural processes underlying auditory scene analysis will be presented, with emphasis on two electrophysiological events thought to index sequential simultaneous grouping processes: the mismatch negativity (MMN) and the object-related negativity (ORN). MMN is a negative component that is elicited by occasional deviant stimuli embedded in a sequence of standard stimuli. It is thought to index an automatic change detection process between the incoming stimulus and a predicted stimulus based on previously presented stimuli. As such, MMN provides information regarding the organization of the stimuli modulate the amplitude of MMN, indicating that it is sensitive to top-down controlled processes. The ORN is a new component that is thought to index the segregation of simultaneous sounds and is present when participants perceive two separate, simultaneous auditory events. Two simultaneous sounds are perceived when one harmonic component of a complex sound is mistuned so that it is not an integer multiple of the fundamental frequency. ORN generation does not depend on contextual factors such as the probability of the mistuned harmonic appearing within a sequence of stimuli and, thus, fundamental frequency. ORN generation does not depend on contextual factors such as the probability of the mistuned harmonic appearing within a sequence of stimuli and, thus, likely reects the discrepancy between the mistuned harmonic and expected frequency based upon the fundamental frequency. Findings from a selective attention task that make strong demands on attentional resources are consistent with the proposal that concurrent sound segregation occurs independently of attentional focus. Our ndings, as well as those from other laboratories, are consistent with an eect of attention on auditory scene analysis and suggest that the processing of sound sequences id dierentially aected by attention than is the processing of sounds that occur simultaneously (i.e., sequential vs. simultaneous grouping processes). The results also indicate that the formation of

5. Training-induced plasticity of the human auditory cortex

Christo Pantev The Rotman Research Institute, Baycrest Centre for Geriatric Care, Toronto, Ont., Canada Perceptual stimuli in dierent modalities reach through the central pathway the corresponding brain neurons, which due to their specic tuning are multiply represented, like a mosaic, in several sensory maps. However, the brain organization is not xed, due to the brains capacity to adapt to current needs of the environment. Several experiments on cerebral cortical organization in musicians demonstrate an astonishing plasticity. Using magnetoencephalographic (MEG) technique in dierent studies, we have investigated the changes that occur in the human cortex if a skill is acquired, such as when learning to play an instrument. We found enlarged cortical representation of tones of the musical scale as compared to pure tones in skilled musicians. Enlargement was correlated with the age at which musicians began to practice. Further, we found that in the left hemisphere the cortical sources for musicians were shifted posteriorly to those of nonmusicians, but not in the right hemisphere. We also investigated cortical representations for notes of dierent timbre (violin and trumpet) and found that they are enhanced in violinists and trumpeters, preferentially for timbres of the instrument on which the musician was trained. MEG measurements were also used to examine the eect of functionally deaerenting the human auditory system. The results suggest that rapid changes can occur in the tuning of neurons in the adult human auditory complex following manipulation of the acoustic environment and a dynamic form of neural plasticity may underlie this phenomenon. In a melody perception experiment, we estimated the involvement of the auditory cortex in the extraction of pitch of complex tones, with special consideration of learning-induced, plastic neural changes. Depending on the prevailing mode of pitch perception, we found signicant dierences in the brain activity in the gamma frequency band. Another important question related to cortical plasticity is the cross-modal reorganization of cortical functions by means of auditory, visual, and somatosensory feedback mechanisms, musicians continually adjust their motor program to t their musical image. For example, trumpet players assess their performance in the auditory modality by listening for errors and in the somatosensory modality by monitoring both the position and the pressure of their lips touching the mouthpiece and their ngers pressing the valves. Errors in musical performance are detected by

130

Rotman Research Institute Abstracts / Brain and Cognition 54 (2004) 128132 patients but lack of available immediate positive reinforcers necessary for rapid learning. Supported by the Deutsche Forschungsgemeinschaft (DFG) and National Institute of Health (NIH).

auditory objects along the horizontal (time) and vertical (frequency) axes recruit distinct neural networks. A neurophysiological model of auditory scene analysis will be outlined at the end of this presentation.

7. Neural correlates of holding information in mind

Mark DEsposito Professor of Neuroscience and Psychology; Director, Henry H. Wheeler, Jr. Brain Imaging Center, Helen Wills Neuroscience Institute and Department of Psychology, University of California, Berkeley, CA, USA To this day, the frontal lobes remain a region of human cortex for which its function continues to elude neuroscientists. Evidence from neuropsychological, electrophysiological, and functional neuroimaging supports the notion that the prefrontal cortex (PFC) is necessary for temporarily maintaining relevant information in an active state, a process that is critical for the voluntary control of behavior. The extensive reciprocal connections from PFC to virtually all cortical and subcortical structures place it in a unique anatomical position to monitor and manipulate diverse cognitive processes. However, little is known about the dierential contribution of PFC versus other cortical/ subcortical areas in implementing executive control. In this talk, I will present evidence from several event-related fMRI studies that support a model of executive control in which PFC biases activity in posterior stimulus-specic association cortex in favor of behaviorally relevant information. Moreover, the temporal dynamics of the signal from the PFC vs. posterior stimulus-specic association cortex is consistent with perceptually driven bottom-up ow of information when encoding representations that must be maintained, and internally driven topdown ow of information when decisions and actions are made based on maintained representations. Hopefully, an improved understanding of the physiological basis of executive control derived from powerful techniques such as fMRI and ERP will lead to a narrower and more useful view of frontal lobe function.

9. Functional neuroimaging of cognition in aging and dementia

Cheryl Grady Senior Scientist, The Rotman Research Institute, Baycrest Centre for Geriatric Care, Ont., Canada; Faculty of Medicine (Psychiatry) and Department of Psychology, University of Toronto, Ont., Canada Although there is currently some debate as to the degree of structural changes in the brain that occur with age, there is little doubt that such changes occur. There are also physiological changes in many areas that could have implications for cognitive function in the elderly. One way to study the impact of these age-related changes in the brain on cognition is to use neuroimaging techniques to examine brain activity during the performance of various tasks and determine how this activity diers between young and older individuals. This approach has been used to study functions such as memory, perception, and attention, and it has generally been found that older individuals utilize dierent areas of the brain than do young subjects when carrying out the same cognitive task. This has led some researchers to suggest that older persons utilize dierent functional brain networks, perhaps to compensate for reductions of eciency in some brain areas. The areas of the brain most often found to be more active during cognitive tasks in the elderly are the frontal lobes. Studies that have directly examined the functional networks utilized during cognition have found that older people do indeed have dierent functional interactions involving the frontal lobes, providing evidence for the idea that they utilize dierent functional networks. In some cases, this dierential activity is directly related to better cognitive performance in the older participants, suggesting that greater reliance on alternative brain regions supports the ability of the older individuals to perform the task. For example, it was found in one experiment that better recognition memory in young adults was associated with correlated activity between the hippocampus and ventral prefrontal areas, whereas in older adults, better memory was associated with functional interactions between the hippocampus and dorsolateral prefrontal cortex. Patients with Alzheimer disease also utilize alternate functional networks during memory tasks, compared to healthy older adults. Evidence that these alternate networks support memory function in the patients has been found and under some conditions these alternate networks involve dorsolateral prefrontal cortex. Thus, both healthy elderly adults and patients in the early stages of dementia show altered cognitive networks that can be used to support memory performance, and specic regions of prefrontal cortex appear to play a crucial role in these compensatory mechanisms.

8. The thought-translation-device (TTD): Taming cognition for action

Niels Birbaumer Professor, Medical Psychology and Behavioral Neurobiology, University of Tubingen, Germany; Adjunct Full Professor of Cognitive Neuroscience, University of Padova/Trento, Italy BrainComputer-Interfaces (BCI) translate single cell or population-recorded neuronal activity into external action, usually computer commands. Animal research using multiple electrode arrays in premotor and motor cortex, particularly by Nicolelis group (Nicolelis, M. (2001). Actions from Thoughts. Nature, 409, 403407) has shown that after training of complex movement patterns, monkeys are able to perform these behaviors directly with brain activity. In humans, the TTD uses non-invasive recording of EEG-patterns, usually slow cortical potentials (SCP) or rolandic mu-rhythms to drive language support programs (LSP). LSPs allow selection of letters and words with the brain signal only in completely paralyzed locked-in patients (Birbaumer et al. (1999). A spelling device for the paralyzed. Nature, 398, 297298). Extensive operant training to self-regulate cortical activity is necessary before restoration of communicative abilities. Operant training of SCP or mu provides patients with auditory and visual feedback and reward of negative and positive SCP usually recorded at the vertex. A new approach using multiple electrodes and forced learning with transcranial magnetic stimulation should improve EEG classication and self-control. fMRI during putamen/pallidum and deactivation of premotor-frontal structures predict learning success. Experiments testing quality of life in the articially respirated and fed completely paralyzed show average quality of life in communicating

10. Understanding the present and plotting the future: Structural and functional brain mapping in stroke and dementia
Sandra E. Black Head, Division of Neurology, Sunnybrook and Womens College Health Sciences Centre; Professor of Medicine (Neurology), University of Toronto, Ont., Canada The advent of Computerized Tomography (CT) in the mid 1970s followed shortly thereafter by magnetic resonance imaging (MRI) in the early 1980s revolutionized clinical Cognitive Neuroscience providing the opportunity for in vivo brainbehaviour correlations as never resting-state patterns of dysfunction in dierent disorders, and activation techniques using the subtraction method with oxygen-15 in the late 1980s took this to another level. Technical advances in the 1990s brought computerized methods for quantitative tissue volumetry of MRI images and functional MRI, which because of the lack of

Rotman Research Institute Abstracts / Brain and Cognition 54 (2004) 128132 radiation and the wider availability, opened the doors to activation studies across the human lifespan and in many structural techniques, which go beyond static cerebral localization to the study of dynamic abilities, at dierent ages and in dierent emerging methodologies for studying brainbehaviour relationships as exemplied by applications in dementia and stroke recovery, using group theoretical modeling. The potential for quantitative structural imaging to serve as surrogate markers of disease progression or retardation will also be considered, as well as the role in dierential diagnosis of topographic patterns of tissue atrophy, which reect the regional selectivity of the common neurodegenerative processes. The possibility of early detection of predementia states using PET, fMRI, and MR spectroscopy will also be discussed. Finally, the ability of fMRI to help to map brain reorganization at dierent stages post stroke, and possibly in response to new interventions will be highlighted. Neuroimaging advances continue to open unparalleled opportunities, but must be disciplined by carefully planned experiments, both functional and structural, with conscientious attention to methodological limitations and quality control, if we are to properly exploit these powerful new methods.

131

11. Functional neuroimaging of autobiographical memory

Brian Levine Scientist, The Rotman Research Institute, Baycrest Centre for Geriatric Care; Assistant Professor, Departments of Medicine (Neurology) and Psychology, University of Toronto, Ont., Canada Autobiographical remembering reects an advanced state of consciousness that mediates awareness of the self as continuous across time. In spite of its importance to adaptive functioning in humans, it has received relatively little scientic attention in comparison to more contrived experiments involving stimuli presented in the laboratory. In the recent years, a handful of functional neuroimaging studies of autobiographical memory have emerged. I will present data from a metaanalysis of this research, with a microcassette recorder. This method allows for the selection of autobiographical events from a pool created earlier by the participant, allowing a high degree of experimental control over autobiographical memory characteristics. We have used this method to study the neural correlates of recollection of everyday episodes that would not normally be accessible using methods of participant-generated autobiographical memory. These ndings will be contrasted to results from recollection of emotional autobiographical events collected on September 11, 2001. The functional neuroimaging data will be interpreted in the context of results from our parallel behavioral studies autobiographical memory in healthy aging, focal brain damage, and dementia, including quantitative analysis of patients high-resolution structural brain images. Our ndings illuminate a distributed fronto-temporo-parietal system mediating autobiographical recollection. They suggest ways in which this system is modulated by memory characteristics such as emotion and rehearsal and dierentially aected by dorsolateral frontal, ventral frontal, and temporal damage.

granted in adults are acquired (an enormous range of perceptual and motor skills, of social skills, expressive and written language(s), understanding and producing music, to name a few). The second is to better understand the neural substrates of these abilities in adults (why are certain regions of the brain most often used for certain types of processing, such as language? What are the correlates predicting poor skill acquisition (i.e., dyslexia), for example). The third is to better understand the full developmental spectrum; learning how, when and in what order brain regions and their associated skills develop, may help in determining the decline in some abilities with aging. In this talk I will review developmental neuroimaging studies that touch on several aspects of cognition, studies which try to disentangle the various processes implicated in tasks, to determine which processes develop when. The studies are largely neurophysiological (event-related potential (ERP) studies), as this is the most readily applied technique with children. I will focus on two areas of research: Visual attention/search paradigms and face processing paradigms. The rst series of studies measured short-latency ERPs and demonstrated the early presence of top-down processing and binding of features in serial presentation tasks. Facilitation of processing with conjunction search is also seen in children, although it continues to develop until teenage years. In more traditional visual search tasks, children show evidence of automatic strategy shifts as a function of task demands, similar to adults, despite the ubiquitous latency and amplitude decreases with age. The results of these studies have important ramications for current models of attentional processing. The second series of studies investigated both early and later stages of face processing in childrenunder conditions of non-directed attention (non-faces were targets), explicit memory tasks and implicit processing of facial emotional expressions. Early stages of processing appear to follow a more protracted developmental course for faces than the simple stimuli in visual search tasks, but the salience of faces is evident at the youngest ages tested. These developmental data make signicant contributions to our understanding of the brain areas involved in perceptual, featural, and emotional processing of faces. Finally, I will present preliminary neuroimaging data from 10 to 12-year-olds in three modalities: functional and structural MRI, and magnetoencephalography (MEG) using a facial expression task, to discuss the convergence of the resultant data and the feasibility of these studies in children.

13. Learning about attentional processes by observing neural activity in extrastriate cortex
Kathleen OCraven fMRI Scientist, The Rotman Research Institute, Baycrest Centre for Geriatric Care, Ont., Canada; Assistant Professor, Department of Psychology, University of Toronto, Ont., Canada Visual attention allows a viewer to select the subset of visual information in a complex scene that is relevant to a particular task or goal and to devote more extensive processing to that information. How does this selection occur? How does it aect visual processing? How can we measure those eects? I will discuss results from one technique: observing the eects of attentional processes on neural activity in extrastriate regions of visual cortex. Certain posterior brain regions are specialized for processing particular aspects of visual stimuli. For example, Area MT/MST responds to visual motion, the fusiform face area (FFA) to faces, the Parahippocampal Place Area (PPA) to spatial layout, and the lateral occipital complex (LOC) to objects. By functionally dening these regions in individual subjects, and then measuring the magnitude of activity in each area during experimental tasks, we can infer the degree of processing for the dierent task conditions (for each attribute). If the same stimulus produces more activity in area MT/MST in one

12. Functional brain measures of cognitive development

Margot J. Taylor Directeur de Recherche, CerCo-CNRS, Universite Paul Sabatier, Toulouse, France Despite the tremendous surge of information from functional neuroimaging in adult populations, much less known about the developing human brain and the underlying neural substrates of cognitive development. Although brain development is most rapid up until two years of age, ongoing structural and functional changes, continue throughout childhooda period of intense cognitive learning and development. There are many reasons to study cognitive development. The rst is to understand how skills that are taken for

132

Rotman Research Institute Abstracts / Brain and Cognition 54 (2004) 128132 cortical region that responds to that attribute (e.g., MT/MST) shows enhanced activity (basic attentional modulation). Additionally, however, the regions which respond to other attributes of the object, which possesses the attended attribute, also show increased activity. While the subject is attending to motion, FFA is more active if the moving item happens to be a face than if a face is the overlapping stationary item. Similarly, PPA is more active if the moving item is a house and the face is the stationary distractor. In each condition, a face, a house, and motion are all present. But the levels of activity within the three corresponding cortical areas depend on what attribute is being attended, and on how the attributes are bound together into objects. Thus, it is possible to determine that attention directed to one attribute of a complex visual stimulus can spread to other attributes of the object, reecting a non-spatial deployment of attention. Current studies are exploring how this object-based attention interacts with spatial attention.

condition than another, for example, this implies motion is being processed to a greater extent in the rst condition. Our group and researchers in other laboratories have used this approach with fMRI data to further our understanding of visual processing. Twenty years ago it was widely believed that the occipital visual areas process visual input veridically, with little top-down inuence. It is now well-established that attention can modulate neural activity in mostif not allsensory areas. Furthermore, we have shown that visual imagery of a particular stimulus recruits the same specialized neural machinery that is recruited to process an actual visual stimulus with the same content. That is, FFA is more active when a subject imagines seeing a face than when she imagines seeing a place, and PPA shows exactly the opposite pattern of results. We can learn more about how attention operates by extending this approach. When two objects overlap in a single location, and subjects attend to one attribute of one of those objects (e.g., motion), the