Fig. 3
Relationship between CBV oscillations, behavior and success rate. fUS imaging is able to differentiate the task response times for fixation/saccade and antisaccade using frequency analysis of CBV changes in the SEF as well as the success rate. Panel a presents an example of two spectrograms related to CBV oscillations in the SEF (large ROI, ~240 pixels) for both animals (S and Y). During the visual task, the CBV in the SEF oscillates around 0.35 Hz and decreases during the rest phase. For each visual task, the fast Fourier transform of the temporal signal in the SEF and the related behavioral signal were plotted and superimposed. A match in peak frequency between the SEF signal and behavior signal indicates synchronization between brain activity and behavior. The behavior signal was created by plotting a sinusoid for which each period was defined by the delay between trial start ti and ti+1, and as shown in Fig. 1c, this signal contains the response time of animals. b Whether in the time domain or frequency domain, there was a set of 13 acquisitions for which a significant correlation between CBV oscillations in the SEF and a related behavior signal was highlighted. c fUS imaging permitted the assessment of the subtle changes in task response time between the saccade and antisaccade experiments (\(\delta t\) = tantisacc – tsacc = 280 ± 210 ms) in agreement with behavior data (\(\delta t\) = 210 ± 160 ms, no significant difference, p = 0.36) using a limited set of 20 successive trials, raw data (gray circles) were summarized with mean (solid red line), 95% confidence interval (gray area), and one standard deviation (white area). d The correlation (level of synchronization) between the SEF signal and behavior signal has been plotted for all acquisitions and was significantly correlated to the success rate of the animal. It was even possible to predict the success rate of the animal only by considering the brain-behavior synchronization over the first 40 s of the experiment. Coefficient correlation and p-value were computed using Pearson correlation test
