Computerized working memory training for hypertensive individuals with executive function impairment: a randomized clinical trial

Abstract

Background: Hypertension is associated with working memory (WM) impairment. However, the benefits of Cogmed WM training for the hypertensive population are unknown. Therefore, we aimed to evaluate Cogmed's effects on the WM performance of hypertensive individuals with executive function (EF) impairment.

Methods: We included 40 hypertensive patients (aged 40-70 years, 68% female) with EF impairment. They were randomized in a 1:1 ratio to receive 10 weeks of adaptive Cogmed training or a non-adaptive control training based on online games. The primary outcome was the WM performance. The secondary outcomes were verbal memory, visuospatial ability, executive function, global cognition, and the neuronal activity measured using functional magnetic resonance imaging (fMRI) under two WM task conditions: low (memorization of 4 spatial locations) and high (memorization of 6 spatial locations). An intention-to-treat (ITT) and per-protocol (PP) analysis were performed.

Results: Cogmed did not show a significant effect on WM or any other cognitive outcome post-training. However, under the WM-low load and WM-high load conditions of the fMRI, respectively, the Cogmed group had an activation decrease in the right superior parietal lobe (ITT and PP analyses) and left inferior frontal lobe (PP analysis) in comparison to the control group.

Conclusion: The Cogmed showed no effects on the WM performance of hypertensive individuals with EF impairment. However, activation decreases were observed in frontoparietal areas related to the WM network, suggesting a more efficient neuronal activity after training.

Keywords: cognitive dysfunction; executive function; hypertension; magnetic resonance imaging; rehabilitation.

Figure 1

Illustration of the WM-low load task and their control condition performed during the functional magnetic resonance imaging exam. (A) WM-low load task cues were presented in a 4 × 4 grid. After a delay, individuals had to indicate if a probe was in the same location that any of the previous cues appeared. (B) Control condition. Individuals were oriented to just look at the image but not memorize them. The numbers below the line represent the display time in ms. Image adapted from Brehmer et al. (2011).

Figure 2

Enrollment of the participants.

Figure 3

Post-training brain activity maps and behavioral performance. (A) In the right superior parietal lobe activation was higher in the control groups than in the Cogmed group after training during the working memory (WM) low-load task condition in the intention-to-treat analysis (ITT) (p < 0.05*). (B) The Same result was observed in per-protocol (PP) analysis. (C) Neuronal activation in the left anterior frontal lobe under the WM-high load condition was higher in the control group compared to the Cogmed group (p < 0.05*), but only in the PP analysis. There was no difference between the groups in behavioral performance (D,E: ITT analysis). *p-values for cluster-wise thresholded z-score equal to or greater than 2.3.