Fixed tests did not confirm a shared association between baseline persistence and late response energy in 9 iEEG and 6 fMRI patients

We tested the relation between regional baseline persistence and late versus early evoked response energy. Rank controls included total response energy and measurement factors. Separate public clinical studies supplied iEEG after amygdala input and BOLD fMRI after local intracranial input. Two iEEG patients supplied development. Fixed tests used 9 other iEEG patients and 6 fMRI patients. Mean person partial rank correlations were -0.0188 and -0.0486. Their 95% person t intervals were [-0.1464, +0.1088] and [-0.1247, +0.0276]. Negative counts were 6/9 and 4/6. Neither fixed shared direction passed. These tests did not establish a cross-modal association. They also did not establish its absence. Shared Holm p values were 0.918972 for the positive claim and 0.742963 for the negative claim. Each claim required a mean and a majority in each modality. An exact stable Gaussian model pair had identical local stationary laws at every lag but different remote response shapes. Energy fractions between model times 1 and 2 were 51.75% and 22.80%. A second construction kept independent white noise and allowed signed couplings. This limits prediction from local data; the models were not fitted to human data. Independent QR fits kept all MRI rank scores. Some nuisance fits were poorly conditioned. Different input sites, short records, uneven regional sampling, surgery, dependent repeats and measurement filters limit the conclusions. We do not identify a common neural cause. Code, derived measurements, source versions, plans, all outcomes, audits, figures and repeat steps are supplied.

human iEEG · fMRI · electrical stimulation · autocorrelation · response energy · local dynamics · identifiability · independent confirmation