Table S5.
Experiment 4: Latency of divergence between conditions, calculated with repeated measures bootstrap resampling (Materials and Methods)
| Cond1 vs. Cond2 | Latency (ms) | ||
| Cond1 | Cond2 | Left hemisphere sensors | Right hemisphere sensors |
| REG5 | RAND20 | 485 (±1) | 396 (±2) |
| REG10 | RAND20 | 691 (±45) | 745 (±28) |
| REG15 | RAND20 | 1,001 (±28) | 1,020 (±19) |
| REG5 | RAND5 | 450 (±1) | 476 (±3) |
| REG10 | RAND10 | 732 (±3) | 862 (±15) |
| REG15 | RAND15 | 1,044 (±19) | 1,192 (±54) |
| RAND5 | RAND20 | 598 (±23) | 400 (±2) |
| RAND10 | RAND20 | 1,621 (±13) | 518 (±8) |
| RAND15 | RAND20 | 1,018 (±22) | 1,016 (±85) |
Because the bootstrap is a stochastic process, slightly different values were obtained for each application of the test. Therefore, for each condition pair, the test was run 1,000 times, and we report the average obtained latency. Numbers in parentheses indicate the SD across the 1,000 runs. We note that the RH latency for RAND10 vs. RAND20 is lower than what might be expected. This effect corresponds to a brief significant interval (not visible in Fig. 6B due to its short duration) that likely stems from noise in the data.