An added test found opposite eye-state coordination contrasts after panel mean removal in 101 EEG volunteers and 47 fMRI adults under Student assumptions

We tested whether eye opening lowers repeatable spatial coordination in two separate human studies. A local plan fixed two 16-site measures. Eight EEG volunteers supplied development. Tests used 101 other EEG volunteers and 47 fMRI adults. Both planned shared-direction tests failed after two-claim Holm correction. After panel mean removal, mean closed-minus-open Q was +0.04369 in EEG and -0.02879 in fMRI. Their 95% person Student limits were [0.03306,0.05433] and [-0.04290,-0.01468]. An MRI interval over every possible eye-state identity matching was [-0.04850,-0.00908], assuming a valid Student model for the true person differences. An added complete direction audit gave opposite-sign Holm p=0.025588 in a ten-claim family. This audit followed MRI outcomes and preceded reserved EEG outcomes. It has no independent MRI confirmation. Broad-signal means were uncertain in both studies. Fixed EEG band controls had opposite signs. An exact, unfitted covariance example also reverses the direction after mean removal while both ranks remain 16. Separate cohorts, EEG baseline order, ocular signals, incomplete MRI motion and processing records, assumed source identities, and unequal signal filters limit neural interpretation. No positive shared direction or common neural cause was confirmed. Original code, all summary facts, source pins and terms, negative tests, figures, checks, and repeat steps are supplied. Source neural arrays remain external.

human neuroscience · EEG · fMRI · eyes open · eyes closed · spatial coordination · cross-validation · subject dependence · reproducibility