The shape and features of the hemodynamic response function (HRF), such as its magnitude, latency and duration, have been reported to carry information on the underlying neuronal activity and physiology. In this study, we explored the regional- and task-specific features of a deconvolved haemodynamic response function (dHRF) across the whole cortex for a wide range of sensorial and cognitive tasks in case of block design stimulations and linear model analyses. To this end, we applied a flexible model based on a sine series expansion of the response function in a cohort of young healthy subjects taken from the Human Connectome Project [1].

The average time courses of the fitted responses and dHRFs are reported in Figure 1. We found an appreciable variation in the BOLD response shape not only across activated regions and tasks, but also across subjects. Quantitative parameters, including amplitude, latency, time-to-peak (TTP) and full width at half maximum (FWHM), are shown in Figure 2. Amplitude and latency showed the highest variability across tasks while time-to-peak and full width at half maximum were relatively more consistent. For all parameters, we observed that the variance across subjects was small compared to the variance across voxels. Our results suggest that the choice of a standard, uniform HRF may be not optimal for all fMRI analyses and may lead to model misspecifications and statistical bias.

References

[1] Van Essen et al., The WU-Minn Human Connectome Project: an overview. Neuroimage, 2013. 80(80): p. 62-79.

Figure 1. Task-specific mean fitted BOLD responses (a) and mean estimated dHRFs (b) averaged in each responding area and then across subjects. In all plots, the average response is represented with a black solid line while the coloured shades around the mean represent the standard error of the mean across subjects.

Figure 2. Across-subjects distribution of amplitude (a) and timing parameters (latency, TTP and FWHM) (b) for each task. Parameters were calculated for each subject using the spatially-averaged task-based response and dHRF.