A comparison of single and multi-echo processing of fMRI data during overt autobiographical recall
Recent years have seen an increase in the use of multi-echo fMRI designs by cognitive neuroscientists. Acquiring multiple echoes allows one to reduce thermal noise and identify nuisance signal components in BOLD data (Kundu et al., 2012). At the same time, multi-echo acquisitions increase data processing complexity and may incur a cost to the temporal and spatial resolution of the acquired data. Here, we re-examine a multi-echo dataset (Gilmore et al., 2021) analyzed using multi-echo ICA (ME-ICA) and focused on hippocampal activity during the overly spoken recall of recent and remote autobiographical memories. The goal of the present series of analyses was to determine if ME-ICA's theoretical denoising benefits might lead to a practical difference in the overall conclusions reached. Compared to single echo data, ME-ICA led to qualitatively different conclusions regarding hippocampal contributions to autobiographical recall: whereas the single echo analysis largely failed to reveal hippocampal activity relative to an active baseline, ME-ICA results supported predictions of the Standard Model of Consolidation and a time limited hippocampal involvement (Alvarez and Squire, 1994). These data provide a practical example of the benefits multi-echo denoising in a naturalistic memory paradigm and demonstrate how they can be used to address long-standing theoretical questions.