Table 1. Summary of internal and external performance of WM, gF, and attention CPMs.
WM CPMs strongly predict HCP 2-back accuracy, HCP PMAT, and memory performance in the SMC dataset. WM CPMs built while controlling for gF at edge selection continue to strongly predict HCP 2-back accuracy. GF CPMs strongly predict HCP PMAT score, HCP 2-back accuracy, and SMC memory performance. When trained on n-back task data, but not rest data, attention CPMs predict HCP SCPT d’. P-values are derived from 1000 permutations of true and null models. Task data is HCP n-back task fMRI data.
| Analysis # | Training behavior, training data | Controlled covariate | Test behavior, test data | Model performance | Validation Type | Data Type | |
|---|---|---|---|---|---|---|---|
| 1 | 2-back accuracy, Task | 2-back accuracy, Task | r = 0.36, P = 1/1001 | * | 10-fold cross-validation | Internal data | |
| 2 | SMC composite memory, Rest | r = 0.37, P = 1/1001 | * | External validation | External data, different behavior | ||
| 3 | PMAT score, Task | r = 0.28, P = 1/1001 | * | 10-fold cross-validation | Internal data, different behavior | ||
| 4 | SCPT d’, Task | r = 0.09, P = 7.19E-2 | 10-fold cross-validation | Internal data, different behavior | |||
| 5 | 2-back accuracy, Task | gF** | 2-back accuracy, Task | r = 0.33, P = 1/1001 | * | 10-fold cross-validation | Internal data |
| 6 | 2-back accuracy, Rest | 2-back accuracy, Rest | r = 0.20, P = 1/1001 | * | 10-fold cross-validation | Internal data | |
| 7 | SMC composite memory, Rest | r = 0.30, P = 1/1001 | External validation | External data, different behavior | |||
| 8 | PMAT score, Rest | r = 0.13, P = 2.00E-3 | 10-fold cross-validation | Internal data, different behavior | |||
| 9 | SCPT d’, Rest | r = 0.06, P = 6.29E-2 | 10-fold cross-validation | Internal data, different behavior | |||
| 10 | 2-back accuracy, Rest | gF** | 2-back accuracy, Rest | r = 0.17, P = 2.00E-3 | * | 10-fold cross-validation | Internal data |
| 11 | PMAT score, Task | PMAT score, Task | r = 0.33, P = 1/1001 | * | 10-fold cross-validation | Internal data | |
| 12 | SMC composite memory, Rest | r = 0.30, P = 1/1001 | * | External validation | External data, different behavior | ||
| 13 | 2-back accuracy, Task | r = 0.34, P = 2.00E-3 | * | 10-fold cross-validation | Internal data, different behavior | ||
| 14 | PMAT score, Rest | PMAT score, Rest | r = 0.12, P = 1.70E-2 | 10-fold cross-validation | Internal data | ||
| 15 | SMC composite memory, Rest | r = 0.36, P = 1/1001 | * | External validation | External data, different behavior | ||
| 16 | 2-back accuracy, Rest | r = 0.13, P = 2.70E-2 | 10-fold cross-validation | Internal data, different behavior | |||
| 17 | SCPT d’, Task | SCPT d’, Task | r = 0.14, P = 5.00E-3 | * | 10-fold cross-validation | Internal data | |
| 18 | 2-back accuracy, Task | r = 0.16, P = 6.39E-2 | 10-fold cross-validation | Internal data, different behavior | |||
| 19 | SCPT d’, Rest | SCPT d’, Rest | r = −5.7E-3, P = 0.533 | 10-fold cross-validation | Internal data | ||
| 20 | 2-back accuracy, Rest | r = 0.11, P = 5.59E-2 | 10-fold cross-validation | Internal data, different behavior |
Correction for multiple comparison: Analyses 1-4, 6-9: alpha= 0.0125; Analyses 5 and 10: alpha=0.05; Analyses 11-13, 14-16: alpha= .016; Analyses 17-18, 19-20: alpha= .025
Significant with P < 0.05, Bonferroni corrected for multiple comparisons within models defined by training behavior and data (Analyses 1-4, 6-9, 11-13, 14-16, 17-18, 19-20).
Additional controlling for gF at edge selection in model construction.