Decoding frequency-specific modulation of fMRI network connectivity: a tACS study

Presented at the 28th Annual Meeting of the Organization for Human Brain Mapping (OHBM 2012). Poster number:  320. 

Earlier studies emphasized regional specificity in cognition and behaviour, while recent studies emphasize the role of connectivity and how it leads to a coordinated interplay between brain regions. Non-invasive brain stimulation, such as transcranial Alternating Current Stimulation (tACS), has been shown to modulate functional connectivity to shape cognitive function. However, it is not clear that different stimulation frequencies will impact brain dynamics differently and that this will be readily measured using imaging approaches that can verify the effects of stimulation. We propose multivariate pattern analyses (MVPA) to demonstrate whether tACS has differentiable effects on regional BOLD activity and connectivity within the brain networks which allow us to decode the stimulation frequency.

Using a randomised epoch-based design, 18 participants underwent four runs of concurrent tACS-fMRI while performing a match-to-sample task. During the task, participants received dual-channel HD-tACS targeting the frontoparietal network (FPN) in the right hemisphere at four pre-specified frequencies (5, 10, 20, 60 Hz). FMRI data were pre-processed and analysed within the Automatic Analysis pipeline, which implemented MVPA decoding using The Decoding Toolbox. We tested the decoder accuracy for both the magnitude of functional activity and the effective connectivity as estimated using psychophysiological interaction (PPI) analysis at voxel, region, and network levels, summarised within the seven networks of the Schaefer-based parcellation.

Univariate analyses confirmed that the task successfully engaged the FPN along with the visual cortex and the basal ganglia network. Decoding based on the PPI connectivity pattern was the most successful in decoding across stimulation conditions; especially within and between the "Dorsal Attention" and "Control" networks, which involve frontoparietal brain regions.

Our study demonstrates that functional connectivity within a task-relevant network is sensitive to specific stimulation frequencies. Our findings provide support for the state-dependency of brain stimulation and pave the way for future closed-loop neuromodulation studies.