Improvement of olfactory fMRI activation and hemodynamic response function curve with respiration correction

Odor perception is highly dependent on respiration, however, the asynchronization between inspiration and odor delivery results in a decreased activation and hemodynamic response function (HRF) curve in olfactory functional MRI (fMRI). This study aimed to investigate the effect of a data post-processing method to minimize the asynchronization and consequently improve the olfactory activation and HRF curve and compare it between different TRs (3 s and 1 s).

A new data post-processing method of respiration correction based on olfactory perception was investigated.

The odor stimuli of olfactory fMRI were mostly based on odor deliveries. But for the reason of respiration, the time of olfactory perception might be a little different. The actual onset time and duration of odor stimuli were modified based on the respiration data.

For 3 s TR data, an increased number of activated voxels and higher HRF curves response height were obtained with the respiration correction method in the primary olfactory cortex (POC) (P = 0.034 and 0.009) and insula (P = 0.024 and 0.001), compared with the uncorrected method. But the activated voxels and response height of the anterior cingulate cortex (ACC) did not differ between uncorrected and respiration correction methods (P = 0.102 and 0.200). The time to peak of the HRF curve was shorter with the respiration correction method in all ROIs (all P < 0.05), compared with the uncorrected method. Whereas there were no significant differences between the uncorrected and corrected results for 1 s TR data.

The post-processing method of respiration correction could effectively minimize the asynchronization between respirations and odor deliveries, and improve the activations and HRF curves for a routine 3 s TR olfactory fMRI data.