Symbolic dynamics applied to the analysis of the effect of breath holding in functional Magnetic Resonance Imaging (fMRI)

Symbolic dynamics was used to investigate the effect of breath-holding on brain functional Magnetic Resonance Imaging (fMRI). Data from an open Internet database were used. Two conditions were compared: breath-holding (on-off task period of 30 seconds) and resting. A symbolic dynamics parameter P(n), equal to the proportion of brain voxels with positive derivatives at each scan in a run, was calculated from the blood oxygenation level dependent (BOLD) signal. The robustness of P(n) was assessed by comparing the results of the original data and data split into odd and even slices. Slice-time correction was also investigated. The average normalized spectrum of P(n) of all breath-bolding runs had greater power peaks at the breath-holding frequency (f(bh) =1/30 Hz), contributing with about 35% of the total signal power, whereas none of the resting runs showed similar behavior. In all cases, P(n) indicated global, periodic effects caused by the respiratory modulation in all breath-holding runs, as opposed to the resting runs.