Tracing the template: Investigating the representation of perceptual relevance
Adaptive perception requires the prioritization of relevant over irrelevant information. When we are looking for a specific book of which we only remember the color of its cover, we can limit our search to mainly that color. The mental representation of what we are looking for is called the attentional template (also target template, search template, attentional set; e.g., Folk et al., 1992). An attentional template is a flexible representation reflecting current selection preferences, as derived from continuously changing task demands and prior selection history. Even though attentional templates are essential for shaping and controlling perception and action in everyday life, surprisingly little is known about their nature. For example, when you look for your car keys, do you look for their shape, their color, or both? In case of the latter, are shape and color integrated in a single representation, or are they independently represented? Can you look for your wallet at the same time, without affecting your key-template? Furthermore, it is often assumed that visual attention is guided by visual templates, but it is perfectly possible that non-visual types of representation (e.g., semantic codes) are also involved. Finally, the nature of the template may change fundamentally in the course of learning, as a result of selection history. The aim of our joint research proposal is to answer the fundamental question of what kind of representation the attentional template is, in terms of function (how it affects our behaviour), physiology (how it is implemented in the brain), and time (how it is affected by learning/selection history). How we flexibly set up new attentional preferences remains one of the great mysteries in cognitive neuroscience, and any reference to templates involves some residual appeal to a homunculus. Although we will not be able to completely banish this homunculus in our research, ourultimate aim is to do the next best thing, to seek fundamental new insights into the nature of this homunculus. To this end, in several other projects we have focused on questions like the number of templates that can be maintained simultaneously, the dynamics of controlling the template, and the influence of different memory systems. In this collaboration we focus on the fundamental question of representation: What is the nature of the attentional template? Specifically, what types of preference can it hold, how do these change as a function of experience, and what are the neural codes underlying these representations? A thorough understanding of the representational properties of attentional templates is a major step on the way towards a neuro cognitive model of attention which will eventually replace the homunculus with a scientific theory of goal-directed perception and action.
in collaboration with:
Prof. Dr. Chris Olivers, Vrije Universiteit Amsterdam, Prof. Dr. Martin Eimer, University College London, Prof. Dr. Hermann J. Müller, Ludwig-Maximilians-Universität München.
Funded by Deutsche Forschungsgemeinschaft (ORAplus program)
Visual learning and attention guidance in patients with macular degeneration
The main problem of eye movement control following central vision loss is that saccades lead to the foveation of peripheral saccade targets. While this is normally adaptive, bringing peripheral points of interest in full view, it is obviously maladaptive after central vision loss, requiring corrective saccades to bring the point of interest into view at a preferred retinal location (PRL) bordering the area of vision loss. What would be more adaptive in this case is to re-reference the saccade target location to an extrafoveal PRL. It is important to note that PRL-use is not the same as saccadic re-referencing to the PRL. In fact, SR has been found to develop only slowly - over months - in clinical populations suffering from foveal vison loss However, recent experiments with central scotoma simulation have demonstrated ways to induce SR over hours rather than months, as reported in the patient studies. While these reports have shown the feasibility of successful SR training with simulated scotomata, they still leave many open questions, as outlined in the work program. Moreover, it took up to 25 hours of training for the fixations with the PRL to become comparably accurate as with the fovea, so even a significant reduction of training hours with improved training techniques would be a considerable progress, making future training programs for patients more feasible. Furthermore, the usefulness of SR-training in AMD-patients (instead of study participants with simulated scotomata) still needs to be established.
Because of the slow spontaneous development of saccadic rereferencing and its importance for efficient visual search (including memory-driven search guidance), the main aim of this proposal is the development of an efficient method to train the fast and durable establishment of saccadic re-referencing (SR) to a PRL in the presence of foveal vision loss and to test effects on memory-guided search in the contextual cueing paradigm as well as its transfer to another important task - reading.
Funded by Deutsche Forschungsgemeinschaft
Neuronal representation of motivational value and context in explicit and implicit learning
We intend to investigate the interaction of the human ventral striatum and medial temporal lobe in complex visual learning. By means of functional imaging we will investigate the interaction of these structures in representing reward expectancy and prediction error in spatial as well as temporal learning contexts. Building on our previous work we will not only investigate actual reward learning, but also reactions to feedback about response correctness as well as the confirmation or violation of implicitly learned contingencies.
This project is part of the collaborative research center SFB 779 http://www.sfb779.de.
Does multivariate pattern analysis of fMRI data profit from high magnetic field strength?
Recently, there were conflicting findings concerning the question whether multivariate pattern analysis (MVPA) benefits from high spatial resolution opf fMRI data. In this project, we intend to systematically investigate the contribution of spatial resolution and sensitivity to MVPA of various tasks at different magnetic field strengths. This is a joint project with the Department of Biomedical magnetic resonance (Prof. Dr. Oliver Speck).
Funded by Deutsche Forschungsgemeinschaft