Abstract #3677

Minimum Resolvable Latency Difference of BOLD Responses at 7T Using Autoregressive Modeling

Santosh Bahadur Katwal¹, James Christopher Gatenby², John C. Gore³, Baxter P. Rogers⁴

¹Electrical Engineering and Computer Science, Vanderbilt University Institute of Imaging Science (VUIIS), Nashville, TN, USA; ²Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science (VUIIS), Nashville, TN, USA; ³Biomedical Engineering, Physics and Astronomy, Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science (VUIIS), Nashville, TN, USA; ⁴Biomedical Engineering, Radiology and Radiological Sciences, Vanderbilt University Institute of Imaging Science (VUIIS), Nashville, TN, USA

Functional MRI (fMRI) at 7T has high signal-to-noise ratio (SNR) and high spatial and temporal resolution, which improve our ability to detect small differences in latency of the BOLD response. The differences in latency can imply causal relationships giving measures of directed influence of one neuronal system on another. We used Multivariate autoregressive (MAR) modeling and Granger causality to determine the directed influence measures and minimum resolvable latency difference of the BOLD response at 7T. Latencies, as low as 112 ms, were resolved in a single subject with just 16 trials. Bootstrapping was performed to obtain 95% confidence intervals on the results.