Figure 1
The race between go and stop and the various ways in which proactive control could improve the success of response inhibition, as indicated by the probability of stopping, p(stop success): (A) base scenario with no proactive adjustments, (B) the effect of slowing down the go response (i.e., the Go RT distribution shifts to the right), and (C) the additional effect of speeding up the stop response (the SSRT distribution shifts to the left). Note that other factors, such as the onset of the go stimulus (Go!) or the delay of the stop signal (Stop!), remained unchanged.
Table 1
List of previous studies that have used stop-trial frequency to examine effects of proactive control on SSRT, plus their main findings.
| STUDY | SAMPLE SIZE | MANIPULATION | MEAN SSRT DIFFERENCE | RESULTS | ||||
|---|---|---|---|---|---|---|---|---|
| AUTHORS | YEAR | HIGH | LOW | REPORTED | EFFECT SIZE3 | TEST | P | |
| Logan & Burkell | 1986 | 12 | 80%, 50% versus 20% | plotted | NA | F(2,44) = 1.34, MSe = 4789 | Insig. | |
| Ramautar et al. | 2004 | 14 | 50% | 20% | plotted | NA | F(1,13) = 0.01 | 0.91 |
| Ramautar et al. | 2006 | 16 | 50% | 20% | –10 msec | –0.25 | F-test | Insig. |
| Verbruggen & Logan (Exp. 5) | 2009a | 18 | 70% | 30% | –13 msec | -1.1 | F < 1 | Insig. |
| Bissett & Logan | 2011 | 24 | 40% | 20% | –19 msec | NA | F(1, 23) = 3.07, MSe = 1395 | <0.10; B = 1.58 |
| Jahfari et al. | 2012 | 16 | 50% | 25% | –19 msec | -0.41 | F-test | Insig. |
| Leunissen et al. | 2016 | 22 | 40% | 20% | –1 msec | -0.03 | t = 0.17 | 0.869 |
| Messel et al. | 2019 | 28 | 66% | 25% | –24 msec | -0.45 | z = –1.62 | 0.106 |
| Bissett et al. | 2021 | 1361 | 40% | 20% | –16 msec | NA | ANOVA | <0.001 |
| Bissett et al. | 2021 | 882 | 40% | 20% | –4 msec | NA | ANOVA | 0.24 |
| Messel et al. | 2021 | 22 | 66% | 25% | –20 msec | -0.39 | z-test | 0.354 |
[i] * B = Bayes Factor.
1 Participants who had short SSDs on average, i.e., mean(SSD) < 300 msec.
2 Participants who had long SSDs on average, i.e., mean(SSD) > 300 msec.
3 Computed by us as Hedge’s g: i.e., the mean SSRT difference divided by the mean standard deviations (i.e., Cohen’s dav; Lakens, 2013) multiplied by an approximation of Hedges’ correction factor (see Goulet-Pelletier & Cousineau, 2018). NA indicate that the effect size could not be computed due to missing information.
Figure 2
Overview of experiment set-ups for Experiment 1 and Experiment 2. (left) The Stop-Signal Task that we used in both experiments; the stop trials were randomly interleaved with go trials and comprised either 50% (high-frequent condition) or 20% (low-frequent condition) of the trials within a block. (right) The sequence of high- and low-frequent blocks; the starting block was counterbalanced between participants.
Table 2
Per experiment, an overview of the go and stop performances for high- versus low-frequent conditions, and the results of the corresponding paired t-test if applicable. Standard deviations are listed inside the parentheses.
| EXPERIMENT 1 (n = 41) | EXPERIMENT 2 (n = 46) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| HIGH | LOW | t(40) | p | dz | HIGH | LOW | t(45) | p | dZ | ||
| Go trials | |||||||||||
| Slowing | 31 (110) | 18 (56) | – | – | – | 7 (67) | 18 (40) | – | – | – | |
| Go RT | 500 (81) | 427 (61) | 10.25 | <0.001 | 1.60 | 521 (112) | 430 (70) | 9.54 | <0.001 | 1.41 | |
| Go omissions | 2.9 (3.0) | 0.8 (1.4) | 4.70 | <0.001 | 0.73 | 2.6 (2.7) | 0.7 (0.9) | 4.88 | <0.001 | 0.72 | |
| Choice errors | 1.3 (1.3) | 2.1 (1.8) | –4.11 | <0.001 | –0.64 | 1.2 (1.2) | 2.0 (1.8) | –3.90 | <0.001 | –0.58 | |
| Stop trials | |||||||||||
| SSD | 260 (83) | 185 (65) | – | – | – | 275 (97) | 182 (74) | – | – | – | |
| p(respond|signal) | 49.1 (4.1) | 51.3 (3.1) | – | – | – | 48.7 (5.0) | 51.1 (2.1) | – | – | – | |
| SRRT | 434 (56) | 389 (50) | 9.28 | <0.001 | 1.45 | 458 (81) | 395 (65) | 11.07 | <0.001 | 1.63 | |
| SSRT | 224 (38) | 235 (41) | –3.40 | 0.002 | –0.53 | 231 (24) | 243 (25) | –3.17 | 0.003 | –0.47 | |
| SSRT per block | 226 (32) | 237 (39) | –3.21 | 0.003 | –0.50 | 231 (24) | 244 (25) | –3.39 | 0.001 | –0.50 | |
[i] dz = Cohen’s d; SSD = stop-signal delay; SRRT = signal-respond reaction time; SSRT = stop-signal reaction time as estimated using the integration approach with replacement of go omissions; SSRT per block = averaged SSRT calculated per block (i.e., to account for proactive slowing).
Note: Accuracy (i.e., go omissions, choice errors; and p(respond|signal)) is reported in percentages; latency (all other variables) is in msec.