Skip to main content
. 2021 Sep 27;11:19147. doi: 10.1038/s41598-021-95884-w

Table 1.

Interactions between day and attentional vulnerability on performance.

Outcome Predictor Interaction By Vulnerability
b (SEM) ΔR2 Level b (SEM) ΔR2
PVT
Lapses§ Day2  − .052*** (.011) .045 LV  − .06 (.07) .001
MV  − .55*** (.07) .101
VOLT
Correct responses Day  − .006 (.007) .001 Mean  − .08** (.03) .010
Hits Day  − .005 (.005) .001 Mean  − .09*** (.02) .021
Correct rejections Day  − .002 (.004)  < .001 Mean .01 (.02) .001
Median RT, correct (ms) Day .429 (.577) .001 Mean  − 15.58*** (2.57) .049
Response confidencea
Incorrect responses Day .046 (.124)  < .001 Mean 1.39** (.49) .014
Misses Day −.107 (.161) .001 Mean 1.57** (.60) .014
False alarms Day .142 (.247) .001 Mean .48 (.95) .001
F2B
Accuracyb Day  − .056** (.020) .011 LV .52*** (.13) .022
MV  < .01 (.13)  < .001
Sensitivityb Day  − .189** (.063) .013 LV 2.22*** (.41) .040
MV .46 (.41) .002
Specificity Day  − .013 (.016) .001 Mean  − .10 (.07) .003
Median RT, hits (ms) Day .969** (.353) .011 LV  − 10.48*** (2.28) .029
MV  − 1.45 (2.28) .001
LOT
Correct responses Day2 .001 (.006)  < .001 Mean .08** (.03) .013
Rotation errorc (M) Day2  < .001 (.001)  < .001 Mean  − .01* (< .01) .009
Excess clicks, correct (M) Day2  − .005** (.001) .023 LV .06*** (.01) .094
MV .02 (.01) .008
Median RT, correct (s)§ Day2  − .004** (.001) .019 LV .02* (.01) .060
MV  − .01 (.01) .011
DSST
Throughput Day2 .015 (.010) .006 Mean .20*** (.04) .031
Median RT, correct (ms)§ Day2  − .190 (.135) .003 Mean  − 2.35*** (.61) .020

Vulnerability calculated as difference between within-person mean of psychomotor vigilance task (PVT) lapses on last sleep restriction day versus last baseline day, per person (see Fig. 6). For significant day or day2*vulnerability interactions, the effect of day or day2 at lower vulnerability (LV, less vulnerable) and at higher vulnerability (MV, more vulnerable) levels is shown. LV estimates were obtained by re-centering attentional vulnerability at 1 standard deviation (SD) below the mean. MV estimates were obtained by re-centering attentional vulnerability at 1 SD above the mean (N = 15 in all analyses; refer to Statistical Analyses). For non-significant interactions, the effect of day or day2 at the mean of vulnerability is depicted.

ΔR2, change in R2 for model term; b, unstandardized beta; DSST, Digit Symbol Substitution Task (processing speed)49; F2B, Fractal 2-Back (working memory)39; LOT, Line Orientation Task (visuospatial processing)48; LV, less vulnerable; M, mean; MV, more vulnerable; PVT, psychomotor vigilance task (attention)66,67; RT, reaction time; s, seconds; SEM, standard error of the mean; VOLT, Visual Object Learning Task (working memory)40.

aRange: 0 (minimally confident) to 100 (maximally confident).

bRange: 0 (minimally accurate/sensitive/specific) to 100 (maximally accurate/sensitive/specific).

cRange: 0 (no rotation error) to 30 (perpendicular line; highest rotation error).

dNumber correct per minute.

§Model includes time of day.

*p < .05; **p < .01; ***p < .001, two-tailed.