Working Memory Performance under a Negative Affect Is More Susceptible to Higher Cognitive Workloads with Different Neural Haemodynamic Correlates

Figures

• 

  Figure 1. The influence of affective states on (a) task accuracy and (b) response time with WML. ANOVA post-hoc analysis of pairwise comparisons was performed using the F-test with significance at * p < 0.05, error bars are S.E.

• 

  Figure 2. Analysis of channel-wise contrast using (a) a paired t-test under merged WML conditions to determine if the mean OxyHb changes towards IAPS images and WM tasks are region-specific and how (b) mean OxyHb and deOxyHb changes with affective states and WMLs during n-back tasks. Intensifying red in (a) indicates a stronger positive contrast of signal amplitude, while intensifying blue indicates negative changes. Pale colour indicates comparable signal level between both periods. Significant activated and deactivated channels were circled with black and grey dotted lines, respectively, after Bonferroni–Holm’s corrections. Warmer colours in (b) indicate higher activation for OxyHb during WM tasks, while colder colours indicate greater reduction in deOxyHb.

• 

  Figure 3. Does haeomodynamic activity vary due to changes in area of activation or strength, and are there any dominant effects neuroanatomically? We plotted (a) the number of significantly activated channels and (b) normalised activation from significantly activated channels against WML and affective states in whole PFC, FP and LRPFC regions. Statistical analysis performed between CG and EG using an independent t-test (2-tailed), with significant * p < 0.05 without family-wise error rate correction and (B) with Bonferroni–Holm’s correction (α = 0.0167). Error bars indicate the standard error (S.E).

• 

  Figure 4. Averaged temporal haemodynamics from significant activated channels in LRPFC during n-back tasks (adapted from [57] with written permission).

• 

  Figure 5. OxyHb temporal properties of significant activated channels in LRPFC during n-back tasks. We plotted the OxyHb signal’s (a) time taken to the highest peak from task onset, and (b) the total area under the curve (AUC) where OxyHb > baseline zero for each affective state and WML. Statistical analysis performed between CG and EG using an independent U-test (2-tailed), with significant * p < 0.05 without multiple-hypothesis test correction and (B) indicates Bonferroni–Holms correction applied (α < 0.0167). Error bars indicate standard error (S.E).