Five of 42 amygdala units fired more during aversive blocks, but patient-level change and firing–high-gamma coupling were not detected

We analyzed the single-unit and macro iEEG recordings in OpenNeuro ds003374 v1.1.1, collected from 9 epilepsy patients during 8 aversive and 9 neutral video blocks of 24 s. Sorted amygdala units were available for 5 patients (42 units; all had mean rates ≥0.1 Hz). Exact two-sided label permutation tests found higher firing in 5 units (12%; 3 patients) and lower firing in 1 unit. The mean patient-level log firing-rate ratio (aversive/neutral) was +0.057 (95% CI −0.090 to 0.195; sign-flip p=0.50; Holm p=1.0). Across blocks, unit rate was not correlated with same-side macro high-gamma power (70–145 Hz; r=0.135, 95% CI −0.25 to 0.49, p=0.63); the condition-adjusted correlation was r=0.131 (p=0.56). High-gamma power was higher in aversive blocks for 4 of 5 patients (+0.043 log10; p=0.125; Holm p=0.50). With five patients, the minimum attainable two-sided exact sign-flip p-value is 0.0625. The group-level results therefore have low power and do not establish absence of an effect. The unit-level count is not a test over independent samples because units share microwires and patients. The findings agree with the source paper only in that a minority of units showed higher firing in aversive blocks; this public subset did not show a consistent patient-level firing change.

amygdala · single-unit activity · intracranial EEG · high-gamma power · aversive stimuli · permutation test · OpenNeuro ds003374