The modulatory effect of adaptive task-switching training on resting-state neural network dynamics in younger and older adults

Figures

• 

  Figure 1. Experimental design of the adaptive task-switching training study27. (a) The structure of the executed tasks. Sitting with eyes closed (1a) and open (1b) for 2 min each, (2) informatively cued task-switching paradigm with no-go stimuli—letter classification and parity task (trained task), (3) non-informatively cued task-switching paradigm with no-go stimuli—letter classification and parity task, (4) informatively cued task-switching paradigm—colour and shape classification, and (5) Attentional Network Test. In the training sessions the difficulty of the informatively cued task-switching paradigm with no-go stimuli was modified block by block based on performance. The resting-state EEG with eyes open from pre- and post-training sessions (1b) were analysed in this study. (b) The schema of the informatively cued task-switching paradigm with no-go stimuli—letter classification and parity task (trained task). TR: task-repeat, TS: task-switching.

• 

  Figure 2. Illustration and summary of the applied analysis methods.

• 

  Figure 3. Flowchart and summary of the EEG data processing steps. ERP: event-related potential, MSE: multiscale entropy, SPD: spectral power density, PLS: partial least squares.

• 

  Figure 4. PLS analysis results for the general aging effect on MSE (using the pre-training resting state in our two age-groups). (Top) The bar graph depicts the contrast between age-groups that was significantly expressed across electrodes and timescales as determined by permutation tests, with error bars denoting 95% confidence intervals. (Bottom) The bootstrap ratio map illustrates the electrodes and timescales at which the contrast displayed in the bar graphs was most stable. Values represent the ratio of the individual electrode weights and the bootstrap-derived standard error (roughly z scores, thresholded at 2.0 which corresponds approximately to p < 0.05). Positive values are plotted in warm clours and indicate timescales and electrodes showing decreases from young to old age-group in resting state MSE. Negative values are plotted in cool colours and denote increases from young to old age-group in resting state MSE.

• 

  Figure 5. PLS analysis results for the general aging effect on SPD (using the pre-training resting state in our two age-groups). (Top) The bar graph depicts the contrast between age-groups that was significantly expressed across electrodes and frequencies as determined by permutation tests, with erros bars denoting 95% confidence intervals. (Bottom) The bootstrap ratio map) illustrates the electrodes and frequencies at which the contrast displayed in the bar graphs was most stable. Values represent the ratio of the individual electrode weights and the bootstrap-derived standard error (roughly z scores, thresholded at 2.0 which corresponds approximately to p < 0.05). Positive values are plotted in warm colours and indicate frequencies and electrodes showing decreases from young to old age-group in resting state SPD. Negative values are plotted in cool colours and denote increases from young to old age-group in resting state SPD.

• 

  Figure 6. PLS analysis results on cue-locked (top row) and target-locked (bottom row) ERP for the group by condition effect (pre- vs. post-training task). On the left side, the bar graph depicts the contrast between experimental groups across age and training conditions that was significantly expressed across electrodes and timepoints as determined by permutation tests, with error bars denoting 95% confidence intervals. On the right side, the ERPs for all of the electrodes used in the original study are shown and the blue circles represent those timepoints where the detected contrast (LV) reaches high significance as determined by bootstrapping. The cue-locked waveforms show the trained task’s ERPs for the young-training group’s pre-training (blue) and post-training (red) as well as the old-training group’s pre-training (magenta) and post-training (green) conditions and the blue circles represent those timepoints where the bootstrap ratio is larger than 4.5. The target-locked waveforms show the trained task’s ERPs for the old-training group’s pre-training (blue) and post-training (red) conditions, and the blue circles represent those timepoints where the bootstrap ratio is larger than 8.1.…

• 

  Figure 7. PLS analysis results for the group by condition effect (pre- vs. post-training resting state) on MSE. (Top) The bar graph depicts the contrast between experimental groups across age and training conditions that was significantly expressed across electrodes and timescales as determined by permutation tests, with error bars denoting 95% confidence intervals. (Bottom) The bootstrap ratio map illustrates the electrodes and timescales at which the contrast displayed in the bar graphs was most stable. Values represent the ratio of the individual electrode weights and the bootstrap-derived standard error (roughly z scores, thresholded at 2.0 which corresponds approximately to p < 0.05). Positive values are plotted in warm colours and indicate timescales and electrodes showing increases from pre- to post-training in resting state MSE. Negative values are plotted in cool colours and denote decreases from pre- to post-training in resting state MSE.

• 

  Figure 8. PLS analysis results for the group by condition effect (pre- vs. post-training resting state) on SPD. (Top) The bar graph depicts the contrast between experimental groups across age and training conditions that was significantly expressed across electrodes and frequencies as determined by permutation tests, with error bars denoting 95% confidence intervals. (Bottom) The bootstrap ratio map illustrates the electrodes and frequencies at which the contrast displayed in the bar graphs was most stable. Values represent the ratio of the individual electrode weights and the bootstrap-derived standard error (roughly z scores, thresholded at 2.0 which corresponds approximately to p < 0.05). Positive values are plotted in warm colours and indicate frequencies and electrodes showing increases from pre- to post-training in resting state SPD. Negative values are plotted in cool colours and denote decreases from pre- to post-training in resting state SPD.