4.3 Latency Variability Within and Across Individuals
Amplitude and latency each provide unique information about brain processing. Amplitude measures are thought to relate to efficiency, resources allocation, capacity, and effort while latency relates to speed of processing (Polich, 2007; Riggins & Scott, 2020; van Dinteren, Arns, Jongsma, & Kessels, 2014). However, studies measuring averaged ERPs often report significant differences in amplitude when comparing task conditions or groups, but few show significant latency differences (Schubert, Löffler, Hagemann, & Sadus, 2022). While healthy adults often show more or less consistent latencies, greater latency variability has been shown in children and individuals with a disability. For instance, decreases in latency variability have been shown to be characteristic of brain maturation (Gavin et al., 2019; Milne, 2011); whereas, when compared to neurotypical participants, greater latency variability have been shown in neurodivergent children (Magnuson et al., 2020), older adults with dementia (Patterson et al., 1988), individuals with schizophrenia (Reinhart et al., 2011; Roth et al., 2007), and individuals who have had a traumatic brain injury (Unsal & Segalowitz, 1995).
Difficulty in obtaining significant mean latency differences between groups may be due to greater measurement error in peak latency measures (Schubert et al., 2022). Peak latencies for most components derived from averaged ERPs have proven to be less reliable than peak amplitudes of the same components in the averaged ERPs and even less so for children compared to adults (Lin et al., 2020; Taylor et al., 2016). This may be because peak latency measures taken from averaged ERPs are not independent from peak amplitude variability in a similar manner as peak amplitude measures from averaged ERPs are not independent of trial-to-trial latency variability (Dupuis et al., 2015; Gavin et al., 2019). That is, in averaged ERPs the latency measures are confounded by fluctuating amplitudes as well as fluctuations of ST latency. An important benefit of using the STP approach is that when the peak amplitude and latency measures are obtained at the trial level for a component, each measure (i.e., amplitude and latency) is independent of the other. Consequently, ST measures of peak latency are shown in this study to be valid and reliable in addition to peak amplitude.