Whole-Brain Task-Based BOLD Anatomical Patterns Hypothesized for the Sternberg Item Recognition Paradigm and their Task-Induced BOLD Changes

Le résumé fourni par la source

Background: Functional magnetic resonance imaging (fMRI) investigations of manipulating information over the short term (working memory; WM) using the Sternberg delayed recognition/delayed recall task have arguably been one of the most active areas of cognitive neuroscience research for several decades. Drawing on a series of fMRI studies spanning 20 years, we have developed an anatomical and temporal account of five task-based brain networks involved in the Sternberg task and how their task-induced blood-oxygen-level-dependent (BOLD) changes are hypothesized to respond to experimental manipulations. Task: The Sternberg Item Recognition Paradigm (SIRP) task put forward for this set of hypotheses involves a string of four or six upper-case consonants being displayed for four seconds, followed by a four-second or zero-second delay. A single probe letter is then shown for two seconds. Participants are asked to respond "yes" or "no" to whether the probe letter presented had been a part of the first string of letters, using a button press with their right hand, with index finger press indicating "yes" and middle finger "no." A fixation cross can be displayed throughout the inter-trial intervals, and to minimize the disparities in basic visual perception between the 4- and 6-letter conditions, pound signs ("#") are included on each end of the 4-letter strings.Hypotheses (anatomical): The hypothesized networks for the Sternberg task can be described as the WM Big 5: Response (RESP), Focus on Visual Features (FoVF), Initiation (INIT), Maintaining Internal Attention (MAIN), and Default Mode Network (DMN). The detailed anatomical depictions of these networks are provided in Tables 1-5.Hypotheses (temporal): Broadly speaking, RESP is expected to peak latest in the trial, and to deactivate mid-trial for long delays. FoVF is expected to deactivate at particular points during the trial. INIT is predicted to be the earliest-peaking network, with a higher peak for the high-load condition. MAIN is expected to peak mid-trial, with a higher peak for the high-load condition, and to initiate activation early-to-mid-trial. The DMN is expected to be mid-trial peaking and show load-dependent deactivation, with the initiation of deactivation coinciding with the trial start. Details are presented in the text, tables and figures below.Conclusions: These anatomical and temporal hypotheses for the WM Big 5 are expected to be upheld over versions of SIRP tasks and samples. However, these networks are also part of a set of networks that are observed across many tasks (zenodo.org/record/4624418), so their functions are not SIRP-specific. Task-general fMRI networks can simplify fMRI investigations, allowing the merging of task fMRI data across multiple sites and SIRP task versions, or even different tasks assessing a range of cognitive domains.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.