Mechanisms of working memory and their modulation by emotion
The aim of this thesis was to examine the possibility that sub-mechanisms of Short-Term or Working Memory (WM) can be selectively affected by task variables such as processing requirements, or situational variables such as positive or negative mood. The model of WM, developed by Baddeley and colleagues (Baddeley and Hitch, 1974; Baddeley, 2000) was taken as the basis for conceptualising the componential architecture of WM. It is proposed that individual WM components, i.e. the executive attention component, the verbal and the spatial storage components, could be systematically modulated by emotions. The first 2 studies investigate the differential involvement of the Central Executive component in tasks that have been reported to be differentially affected by mood. It has been argued that analytical tasks make greater demands on the Central Executive (CE) and therefore alternative demands on the CE will disrupt such tasks. The detrimental effect of positive mood on analytical tasks is due to the disruptive effect of this type of mood on CE function. Using a dual-task paradigm, it was demonstrated that an analytical reasoning task was more disrupted by a concurrent WM task than was a creative task, i.e. a demonstration that the CE is more involved in analytical tasks. Measures of brain electrical potentials demonstrated a parallel effect at the cortical level. The following studies examined other WM components, the verbal and spatial "slave" mechanisms and the possibility that they may be differentially modulated by emotion. Recent behavioural, psychophysiological and neuroimaging evidence suggests that spatial but not verbal, WM system is largely reliant on visual attention. It has also been shown that certain emotions, i.e. anxiety or fear, may be intrinsically associated with visual attention. On the basis of this overlap, it was predicted that induced anxiety/ fear would selectively modulate spatial WM, with little effect on verbal WM performance. An n-back task allowing of fine-grained dissociation of verbal and spatial WM was developed and three studies examined the effects of threat of shock on verbal and spatial n-back WM performance. The studies confirmed the prediction, showing that spatial WM performance was impaired in conditions of threat, while verbal WM performance remained unchanged. A number of analyses emphasised the relationship between anxiety and the decrement in spatial n-back performance. The presence of anxiety was objectively validated via physiological recordings. The literature contains suggestions that, in contrast to the selective disruption of spatial WM in threat-anxiety observed in our studies, trait and test anxiety tend to affect verbal WM more than spatial WM. Therefore, in a subsequent study verbal and spatial n-back WM performance was contrasted in high vs. low trait anxiety individuals. It was found that verbal and spatial WM were not differentially disrupted in individuals high with trait anxiety. The final study aimed to determine whether interhemispheric asymmetry models in emotion and WM could serve as a neuroanatomical explanation of the selective effect of anxiety on spatial WM. The study employed neuroimaging technology to compare the activation of regions of the frontal cortex in the two hemispheres in the verbal and spatial n-back tasks employed earlier. The anatomical regions of interest were chosen on the basis of neuroanatomical models of brain laterality in emotion. The results do not contain reliable patterns of frontallateralization. Hence, they provide little support to laterality-based accounts of the interaction between spatial WM and anxiety. These findings support the hypothesis that WM components can be selectively modulated by factors such as affective state. They also suggest that attentional-executive mechanisms may not be fully distinct from the storage systems, as claimed in the original characterisation of WM.