Neuronal correlates of focus updating in working memory
Final Report Abstract
Working memory (WM) enables the retention of a limited number of items for a short period of time. In addition, many situations require that a subset of WM contents be transiently prioritized for processing by focusing attention on them. Current models of WM suggest that the same attentional mechanisms that are known from perception also operate in WM. In vision, shifts of attention between spatial positions within the boundaries of one object are performed faster than shifts between positions located on different objects. This within-object benefit can be explained by an automatic spread of attention within perceived object boundaries in visual cortex. Hypothesizing the same attentional mechanisms in WM as in perception, we tested whether the within-object benefit can be observed, both on the behavioral and neural level, when subjects focus attention on spatial positions that are no longer physically present but represented in WM. We found that attentional shifts in WM were faster for spatial positions located on the same object compared with equidistant positions on separate objects. Shifting attention in WM thus showed a within-object benefit comparable to the effect observed in a perceptual version of the same task. This behavioral benefit was associated with increased BOLD activity in posterior parietal cortex that could not be explained by differences in eye movements between conditions. Moreover, analysis of retinotopic visual cortex revealed that the automatic spread of attention within object boundaries was present also for information held in WM. Specifically, when attention was shifted to a memorized position, activity in early visual areas was enhanced at the retinotopic location corresponding to the second position co-located on the same object compared to an equidistant position located on the other object. These results extend the previous suggestion of shared mechanisms of spatial attention in perception and WM by demonstrating that this notion also holds for object-based attention. The study was selected for a Goethe University press release (http://www.goethe-university-frankfurt.de/54600288/031). In our first new project we investigated the neural bases of capacity limitations in WM by combining fMRI with cognitive modeling of behavioral performance. We provided evidence that the superior intraparietal sulcus (IPS) is a critical brain region that influences the variability of WM precision between and within individuals under increased memory load, possibly in cooperation with perceptual systems of the occipital cortex. These findings underline the importance of integrating cognitive modeling with univariate and multivariate methods in fMRI research to improve our knowledge of brain-behavior relationships. This work was done in close cooperation with Prof. Fiebach's laboratory. In a second new project we developed and tested a novel approach to estimate the memory precision of a single item within a trial. We asked subjects to successively report each of the previously encoded orientations from WM in a continuous recall format. A sequence of behavioral experiments was employed to provide a detailed characterization of the obtained information. We found that subjects were able to recall information about all items above chance even at set size six. Precision, however, varied systematically with report order and set size. Specifically, the first item in a recall sequence was reported with the highest precision that was comparable to the typical situation when only one item was to be reported. Subsequently reported items were reproduced at increasingly reduced precision, with the reduction in precision from the first to the second recalled item being clearly steeper than subsequent declines in the sequence. This special status of the first recalled item can be eliminated by attended visual interference before the recall of the first item but not by a longer delay. These results show that the whole-report procedure is feasible, and yields information on the within-trial differences of item precision that is not available from standard single report paradigms. By accounting for the interference of the whole-report sequence, we were able to separate the distribution of within-trial information from across-trial performance fluctuations. This in turn might improve the identification of the neuronal processes underlying limitations in WM by directly relating the item-wise estimated parameters to concurrently measured brain activity using MEG signals. The MEG data have been recorded and are currently being analyzed.
Publications
- (2015). Activity in human visual and parietal cortex reveals object-based attention in working memory. Journal of Neuroscience 35, 3360-9
Peters B., Kaiser J., Rahm B., Bledowski C.
(See online at https://doi.org/10.1523/JNEUROSCI.3795-14.2015) - (2016). Superior intraparietal sulcus controls the variability of visual working memory precision. Journal of Neuroscience 18, 5623–35
Galeano Weber E.M., Peters B., Hahn T., Bledowski C., Fiebach C.J.
(See online at https://doi.org/10.1523/JNEUROSCI.1596-15.2016)
