Spatially distributed effects of mental exhaustion on resting-state FMRI networks

Abstract

Brain activity during rest is spatially coherent over functional connectivity networks called resting-state networks. In resting-state functional magnetic resonance imaging, independent component analysis yields spatially distributed network representations reflecting distinct mental processes, such as intrinsic (default) or extrinsic (executive) attention, and sensory inhibition or excitation. These aspects can be related to different treatments or subjective experiences. Among these, exhaustion is a common psychological state induced by prolonged mental performance. Using repeated functional magnetic resonance imaging sessions and spatial independent component analysis, we explored the effect of several hours of sustained cognitive performances on the resting human brain. Resting-state functional magnetic resonance imaging was performed on the same healthy volunteers in two days, with and without, and before, during and after, an intensive psychological treatment (skill training and sustained practice with a flight simulator). After each scan, subjects rated their level of exhaustion and performed an N-back task to evaluate eventual decrease in cognitive performance. Spatial maps of selected resting-state network components were statistically evaluated across time points to detect possible changes induced by the sustained mental performance. The intensive treatment had a significant effect on exhaustion and effort ratings, but no effects on N-back performances. Significant changes in the most exhausted state were observed in the early visual processing and the anterior default mode networks (enhancement) and in the fronto-parietal executive networks (suppression), suggesting that mental exhaustion is associated with a more idling brain state and that internal attention processes are facilitated to the detriment of more extrinsic processes. The described application may inspire future indicators of the level of fatigue in the neural attention system.

Figure 1. The research design consisted of two days.

On the free day, participants were in the lab only in the morning and in the afternoon. In between, they performed self-chosen, low-effort activities. On the working day, morning and afternoon schedules were kept identical, however, an additional resting state measurement was included in the noon. Time scale is not proportional see methods for timing information. RS  =  resting state; FT  =  flight task.

Figure 2. Exhaustion levels over the course of the work (wd_t1/t3) and the free day (fd_t1/t3) with standard error bars.

Figure 3. Effort ratings for all N-back conditions as an effect of treatment (work vs free days): Estimated marginal means per time-of-day and day activity (value of the centered covariate day order  = 0).

Error bars represent standard errors.

Figure 4. Group ICA results for the analysis of the pooled baseline resting state scans (WD-T1, FD-T1): (a) anterior default mode network; (b) posterior default mode network; (c) visual network; (d) sensory-motor network; (e) right fronto-parietal network; (f) auditory network; (g) left fronto-parietal network.

Statistical thresholds were corrected for multiple comparisons.

Figure 5. Early visual processing network.

(a) Statistical map of the differential effects (WD_T3>WD_T1+FD_T1+FD_T3). Statistical thresholds were corrected for multiple comparisons. (b) Bar graph of the regional ICA scores (with standard error bars) in all conditions. (c) Correlation graph of the regional ICA scores against the normalized degree of exhaustion for the separate working day sessions (with fit lines indicating the directions of the correlations).

Figure 6. Anterior default mode network.

(a) Statistical map of the differential effects (WD_T3>WD_T1+FD_T1+FD_T3). Statistical thresholds were corrected for multiple comparisons. (b) Bar graph of the regional ICA scores (with standard error bars) in all conditions. (c) Correlation graph of the regional ICA scores against the normalized degree of exhaustion for the separate working day sessions (with fit lines indicating the directions of the correlations).

Figure 7. Left (a,b) and right (c,d) fronto-parietal networks.

(a,c) Statistical maps of the differential effects (WD_T3>WD_T1+FD_T1+FD_T3). Statistical thresholds were corrected for multiple comparisons. (b,d) Bar graphs of the regional ICA scores (with standard error bars) in all conditions.