Temporo-Spatial Characteristics in Working Memory Processes Investigated with Static and Dynamic Complex Stimuli - Functional Magnetic Resonance Imaging, Event-Related Potentials and fMRI-Constrained Source Analysis

Experiment 1:In the majority of experiments studying working memory (WM) tasks either delayed match-to-sample (DMTS) or n-back tasks were used to assess WM operations. In n-back tasks, however, a separation of different WM processes is impossible, whereas in DMTS tasks it is usually not required to monitor the WM content.In Experiment 1, a variant of a DMTS task with continuously morphing stimuli was applied, requiring human participants to continuously monitor and compare the currently presented stimulus to the held-in-mind target shape during the retention period. To asses changes caused by cognitive load two different types of stimuli were employed: Complex curved shapes and a simple circle. Furthermore four different delay durations (3, 6, 9, 12 s) have been applied. The goal of the present fMRI study was to characterize brain areas responsible for WM maintenance under continuous monitoring conditions in a WM task with morphing shapes. Moreover, the influence of cognitive load on maintenance activity in motion-sensitive area MT should be investigated. Data from 15 study participants are presented. Conjunction analysis including the contrasts complex versus simple shapes over all delay durations revealed activations in a widespread fronto-parietal network, which has been discussed to be involved in WM by a large body of literature. Area MT showed an association with target complexity, resulting in a significant signal enhancement while monitoring complex shapes (compared with simple circle trials).The present results demonstrate that the level of activation during the maintenance interval might be dependent on target complexity, because complex targets caused a stronger activation than simple circle targets in brain areas associated with WM processes even during similar visual stimulation.Experiment 2:So far, WM retrieval has been investigated only in a few studies. In Experiment 2, a DMTS task using static stimuli was applied to examine differences in the processing of different probe types (targets, non-targets) during the retrieval epoch. Probe stimuli were presented at different positions in the trial (trial position A, B, and C), resulting in parametrically increasing delay durations (3, 7, and 11 seconds) before probe presentation.Twenty individuals completed a similar WM task in an fMRI session as well as in an EEG session.The objective of the present study was to characterize differences in the cognitive processing of target probes and non-target probes by using different methodological approaches (behavioral, fMRI, EEG, and SA data) and to examine the influence of delay duration (cognitive load) on these differences.P3b mean amplitudes showed higher values for target trials compared with non-target trials at electrode position Cz, corroborating old/new effects reported from long-term memory research.The majority of brain regions which showed activation in fMRI in the probe epoch regardless of probe type (probe epoch versus fixation) did also show activation when contrasting target trials with non-target trials. This indicated a specific involvement of these brain areas in the processing of target stimuli. Altogether, behavioral, fMRI and EEG data indicated probe type-related processing differences, which might have been influenced by the ratio of target stimuli to non-target stimuli and also by the presence of distracting probe stimuli during longer delay periods (7, 11 s).Source analysis revealed a sequence of source activities during probe type processing starting with activity in occipital and temporal brain regions. This was presumably linked to the processing of stimulus features, followed by a simultaneous involvement of parietal and frontal brain regions and later processing in superior frontal gyrus (pre-SMA).Differences in source activity between targets and non-targets indicated a specific involvement of left fusiform gyrus in the non-target condition, probably associated with the mental imagination of the target stimulus during non-target probe processing.Furthermore, source activities showed specific engagements in target processing for the regional source in anterior cingulate cortex (ACC) before response execution and also for the regional source in superior frontal gyrus (SFG) before and simultaneously to response execution. These findings might be an indicator for the involvement of both regions in different stages of conflict managing operations because target trials had a lower stimulus frequency compared with non-target trials (at trial positions A and B).Summarized, different WM processes (maintenance, retrieval) were investigated in both experiments, and different methodological approaches were applied. In the general discussion, findings from both experiments are linked to each other, considering brain regions which were involved in the processing of both tasks