Delay discounting is predicted by scale-free dynamics of default mode network and salience network

- We explored the effect of fractal properties of neural networks on delay discounting.
- We conduct Independence Component Analysis (ICA) to extract neural network.
- Delay discounting positively correlated with Hurst exponent of DMN and SN.
- Scale-free dynamics properties of DMN and SN play a crucial role on delay discounting.

Resting-state functional Magnetic Resonance Imaging (rs-fMRI) is frequently used as a powerful technology to detect individual differences in many cognitive functions. Recently, some studies have explored the association between scale-free dynamic properties of resting-state brain activation and individual personality traits. However, little is known about whether the scale-free dynamics of resting-state function networks is associated with delay discounting. To address this question, we calculated the Hurst exponent which quantifies long-term memory of the time series in resting-state networks (RSNs) identified via independent component analysis (ICA) and examined what relationship between delay discounting and the Hurst exponent of RSNs is. ICA results showed that entire nine RSNs were successfully recognized and extracted from independent components. After controlling some covariates, including gender, age, education, personality and trait anxiety, partial correlation analysis revealed that the Hurst exponent in default mode network (DMN) and salience network (SN) was positively correlated with the delay discounting rates. No significant correlation between delay discounting and mean Hurst exponent of the whole brain was found. Thus, our results suggest the individual delay discounting is associated with the dynamics of inner-network interactions in the DMN and SN, and highlight the crucial role of scale-free dynamic properties of function networks on intertemporal choice.