Relating Unpredictable Expectancies in Working Memory to Complex Tasks: Network Connectivity and Training-Related Transfer
O'Connell, Margaret Anne
MetadataShow full item record
Complex tasks involve competing elements, have subtasks and/or varying information loads that involve working memory, include unspecified strategies for task completion, and are required when successfully navigating challenges in everyday life. A strong relationship between working memory and complex tasks has prompted the use of working memory training to improve complex task performance. However, these past interventions have resulted in mixed findings. In the current dissertation we investigate whether these mixed findings are driven by the variability in probe-cue expectancies used in these intervention studies. Therefore, in the two experiments presented in this dissertation, we investigated the relationship between complex tasks and the probe-cue expectancies in working memory. In the first experiment, a randomized control trial, we compared potential improvements from two types of working memory updating training, one with predictable probe-cue expectancies, and other with unpredictable probe-cue expectancies, against an active control group. All three groups were assessed on complex task performance before and after 10 hr (across 5 sessions) of either training or completion of 10 hr of word puzzles (active control group). The two working memory training groups showed significantly more improvements in a complex task outcome than the control group. The complex task outcome included standardized tasks of episodic memory, reasoning, task switching, and inhibition. When these cognitive constructs were dichotomized, we found that the two working memory training groups improved significantly more than the control group in both the inhibition construct and the episodic memory construct. Using performance in the trained tasks, we investigated the effect of individual differences on transfer, finding that initially high performers and better learners in the unpredictable expectancies group had significant improvements in the complex task, inhibition, and episodic memory measures, compared to the control. However, only low initial performers in the unpredictable expectancies group, and not the predictable expectancies group, resulted in these same improvements for the inhibition measure, compared to the control group. In the second experiment, we investigated how skill learning of novel complex tasks are related to task-related functional connectivity of two attentional networks (cingulo-opercular, CO; fronto-parietal, FP). CO and FP seed-to-voxel connectivity was investigated during three tasks of increasing cognitive control demands; two of which included unpredictable probe-cue expectancies. Complex task learning was related to within-network connectivity for these attentional networks (e.g., CO to CO), and to between network connectivity for the attentional networks (e.g., FP to CO) and to the default mode (DM) network (e.g., FP to DM) during the tasks with unpredictable probe-cue expectancies. These two studies indicate that better learners in complex tasks utilize within- and between-network connectivity of the neural networks that underlie working memory during tasks with unpredictable expectancies. Furthermore, training in working memory, particularly with unpredictable probe expectancies, will transfer to complex tasks. These findings suggest why working memory and complex tasks are related (Experiment 2) and how we can strengthen this relationship (Experiment 1).