Fig 1.
Illustration of the letter detection task during low load (top) and high load (bottom).
Subjects had to respond when the letter ring included an X (target). Task-irrelevant complex tones (standard and deviants) were presented simultaneously with the onset of the letter rings. Stn = standard, Dev = deviant.
Table 1.
Means (and SDs) for behavioral performance and mean amplitudes for the four conditions (deviance by load).
Table 2.
Results of repeated-measures ANOVAs with the independent variables Deviance (deviant minus standard), Load (low load minus high load), and their interaction on behavioral performance and mean amplitudes.
Fig 2.
Grand mean (N = 28) ERP waves (referenced to tip of nose) after tone onset for the relevant electrodes, separately for the four conditions (Deviance x Load).
For visualization purposes only, the ERPs were low-pass filtered (30 Hz, 6 dB roll-off, Butterworth). The interval between 160 and 220 ms after tone onset was used to capture the mismatch negativity.
Fig 3.
Grand mean (N = 28) ERP difference waves (deviant minus standard, referenced to tip of nose) after tone onset at Fz (top), Cz (middle), and mastoids (bottom) for low and high perceptual load (left), the subsequent difference of low load minus high load (middle), and a scatterplot of individual mean MMN amplitudes (i.e., deviant minus standard) for low and high load (right).
For visualization purposes only, the ERPs were low-pass filtered (30 Hz, 6 dB roll-off, Butterworth).
Fig 4.
Mean amplitudes (referenced to tip of nose) for the four conditions (Deviance x Load) across 160–220 ms after tone onset at Fz (left) and Cz (right).
The top row shows condition means (95% CI for each variable), and the bottom row shows mean difference scores (95% CI for the differences) between conditions for effects of load (low minus high), tone deviance (deviant minus standard), and the interaction (i.e., [deviant–standard at low load]–[deviant–standard at high load]).
Fig 5.
Meta-analysis of how visual task demands changed the MMN (forest plot on left and funnel plot on right).
The analysis focused on the mean amplitudes at Fz (between about 100 and 250 ms after stimulus onset) and was expressed in μV (with 95% CI). The overall negative effect shows that the MMN (i.e., deviant minus standard) was more negative during low than high visual demands (i.e., MMN during low demands minus MMN during high demands). See main text for references to the individual studies. In the forest plot (left), the size of each square corresponds to the weight in the meta-analysis.
Table 3.
Statistics for amplitudes at Fz used in the meta-analysis.