Contributed Talk Sessions | Poster Sessions | All Posters | Search Papers

Poster Session B: Wednesday, August 13, 1:00 – 4:00 pm, de Brug & E‑Hall

Fast fMRI signals up to 1Hz vary across brain states and predict spontaneous neural activity

Leandro P. L. Jacob¹, Sydney M. Bailes, Carsen Stringer², Jonathan R. Polimeni, Laura D. Lewis; ¹Massachusetts Institute of Technology, ²HHMI Janelia Research Campus

Presenter: Leandro P. L. Jacob

fMRI signals were traditionally seen as very slow, with most resting-state studies only investigating signals slower than 0.1Hz, but task-based studies have shown that fMRI signals up to 0.75Hz reflect stimulus-induced neural activity. Here, we investigate whether high-frequency fMRI signals can index spontaneous neural activity across brain states. Using simultaneous EEG-fMRI in 21 humans drifting between sleep and wakefulness, we found an increase in fMRI spectral power during NREM sleep (compared to wakefulness) across frequency ranges as fast as 1Hz. Using machine learning, we found that these fast fMRI signals predict fluctuations in canonical neural rhythms measured with EEG, in subjects held-out from the training set. Since fMRI signals and neural rhythms are sensitive to systemic physiology, we tested whether this predictive fast fMRI information specifically represented neurovascular coupling, or was also present in the ventricles. We found that fMRI signals as fast as 0.9Hz (for alpha rhythm predictions) and 0.8Hz (for delta rhythm predictions) contained unique neural information above what was present in the ventricles. These results reveal that high-frequency spontaneous fMRI signals are coupled to neural activity that varies across brain states and index cognitive processes, pushing the boundaries of fMRI’s abilities to reveal brain dynamics underlying cognition.

Topic Area: Brain Networks & Neural Dynamics

Extended Abstract: Full Text PDF