Block 1 · block_h.csv
120 trials · accuracy
1.7% · mean RT
0.602 s
Alerting — · Orienting — · Conflict -235 ms
NC 0.719 s · CC — · DC — · SC 0.484 s
Computed on correct trials only across 1 block(s); positive scores are interpreted as the network being engaged (slower RT in the harder condition). The ± figure is the standard error of the mean difference; the 95% CI uses the normal approximation. For Bayesian credible intervals (CrI) updated against literature priors, see the Bayesian tab.
| Block | Trials | Accuracy | Mean RT | Alerting | Orienting | Conflict |
|---|---|---|---|---|---|---|
| Block 1 block_h.csv | 120 | 1.7% | 0.602 s | — | — | -235 ms |
| Aggregate | 120 | 1.7% | 0.602 s | — | — | — |
| Cue | Mean RT |
|---|---|
| NC No cue | 0.719 s |
| CC Centre cue | — |
| DC Double cue | — |
| SC Spatial cue | 0.484 s |
| Congruency | Mean RT |
|---|---|
| congruent | 0.719 s |
| incongruent | 0.484 s |
| neutral | — |
Histogram + Gaussian KDE per cue type, correct trials only. Click a legend entry to hide that series.
Same data grouped by the central-arrow's flanker congruency. Neutral trials use the dashed neutral target.
Closed-form Normal–Normal conjugate update:
prior on the participant's mean μ for each ANT score, then per-block
scores treated as observations x_i ~ Normal(μ, σ_L) with
σ_L = 100 ms. The posterior is analytic, no MCMC needed.
Priors come from the per-block Turing model in the previous Julia
analysis (centred on Fan&Posner-style population means with σ=100 ms,
deliberately broad). Reference column is the population distribution
from Fan & Posner.
| Score | Blocks (n) | Prior μ ± σ (ms) | Reference μ ± σ (ms) | Posterior μ ± σ (ms) | 95% CrI for μ |
|---|---|---|---|---|---|
| Alerting | 0 | 40 ± 100 | 47 ± 18 | 40.0 ± 100.0 | [-156.0, 236.0] |
| Orienting | 0 | 50 ± 100 | 51 ± 21 | 50.0 ± 100.0 | [-146.0, 246.0] |
| Conflict | 1 | 98 ± 100 | 84 ± 25 | -68.6 ± 70.7 | [-207.1, 70.0] |
Caveat: with only one block of data, the per-score posterior is dominated by the prior. Run more blocks to let the data update beliefs more strongly.