Introduction: Sustained activity in frontal and parietal brain areas during spatial working memory is thought to reflect active storage and maintenance processes, but the nature of the representations carried by this activity continues to be investigated. The aim of the doctoral thesis was to investigate neural mechanisms of spatial representations in working memory and how these representations change depending on task demands and in relation to individual differences. First, we were interested in which type of positional information, expressed in different coordinate systems, is represented in neural activity during working memory. Second, we were interested in strategies used to encode and maintain spatial representations and the role of individual brain areas in spatial working memory processes.
Methods: We conducted two separate studies. In Study I, we combined data from six experiments, including a total of 158 participants (78 women, 25 ± 6 years). In Study II, we conducted two experiments with 31 (20 women, 23 ± 5 years) and 44 (23 women, 21 ± 2 years) participants, respectively. In both studies, participants performed a spatial working memory task, where they were asked to remember the position of a briefly presented target stimulus and, after a delay period, to move a disk using a joystick to the position of the remembered target, while we measured their brain activity using fMRI. In Study II, different task conditions were included, designed to manipulate the use of retrospective sensory or prospective motor coding of spatial representations.
Results: In Study I, our results revealed no significant relationships between brain activity and any of the behavioral measures of memory precision, which would indicate the type of positional information encoded and maintained in spatial working memory. However, we observed systematic biases in behavioral performance that are thought to reflect the use of categorical spatial representations. The results showed that participants who strongly relied on categorical representations had poorer fine-grained memory precision, which was reflected also in the brain activity during the task. In Study II, prospective motor coding was associated with increased activity and functional connectivity in somatomotor areas, whereas retrospective sensory coding was related to increased activity and functional connectivity in parietal and visual areas.
Conclusion: Our findings contribute to the understanding that multiple representations and strategies can be used to encode and maintain information in working memory. They also suggest that the available strategies are not mutually exclusive and that their use may depend on task demands and individual differences.
|