Temporal stability of auditory event-related potentials in healthy individuals and patients with traumatic brain injury.

Historically, cognitive event-related potentials (ERPs) have received limited acceptance for clinical use due to lack of evidence for their reliability. However, recent advances in computer technology and artifact rejection methods have greatly enhanced the fidelity of ERP measurements. The present study examined the test-retest reliability of ERP measurement by using current data processing methods. We assessed the temporal stability of the most commonly used ERP paradigm, auditory pure-tone "odd ball" detection, and compared it with other commonly used clinical measures reported in the literature. Auditory ERPs were collected in 19 healthy subjects and 7 patients with traumatic brain injury at two time points, 2 days to 2 months apart. Test-retest reliability was calculated for four ERP components: N1, MMN (mismatch negativity), P3, and N4. In healthy subjects, temporal stabilities of these four commonly studied ERP components' amplitude measurements were moderate to high, with intraclass correlations ranging from 0.6 to 0.8. In contrast, in patients with traumatic brain injury, ERPs were stable only for the N1 component (intraclass correlation = 0.7).