Poster C102, Sunday, March 26, 5:00 – 7:00 pm, Pacific Concourse
Measuring the impact of short-term training on brain networks using resting state connectivity
Adam Steel1,2, Cibu Thomas1, Aaron Trefler1, Gang Chen3, Chris Baker1; 1Laboratory of Brain and Cognition, National Institutes of Health, 2Oxford Centre for Functional MRI of the Brain, University of Oxford, 3Statistics and Computing Core, National Institutes of Health
While training has been shown to alter resting-state connectivity, the extent to which such effects are influenced by other factors (e.g. diurnal changes) is unclear. Here, we address these issues using a within-subject design involving 21 healthy adults over four experimental visits. Each visit included two scan sessions, at 1000 and 1400 hours (AM/PM). On two visits, between the AM and PM scans participants trained for 90 minutes on either a visuospatial-learning (VSL) or a motor sequence-learning (MSL) task. On the remaining two visits, participants received no training to test for diurnal changes in connectivity. We focused on connectivity strength with the bilateral hippocampi (HIPP) and motor cortices (MC), given their involvement in VSL and MSL, respectively. Both MC and HIPP networks showed diurnal fluctuations in connectivity. After controlling for diurnal fluctuations, the effect of VSL on HIPP was evidenced by a training-related decrease in connectivity between the hippocampi and sensorimotor-, premotor-, and lateral prefrontal cortex, but an increase in connectivity with anterior medial place area and putamen. Pre-training connectivity between HIPP and ventral prefrontal cortex predicted VSL. In contrast, MSL evoked an increase in connectivity between MC and temporal and occipitotemporal cortices, but a decrease in connectivity between MC and the caudate nucleus. Better performance in MSL was positively correlated with an increase in connectivity between MC and putamen and cerebellum. These results suggest that resting-state fMRI can detect task-specific effects of training, but also underscore the importance of controlling for potential confounds in connectivity analyses.
Topic Area: LONG-TERM MEMORY: Skill learning