The Neuronal Mechanism Of Temporal Discounting:Based On The Complex Network Algorithms

Delay discounting generally reflect the scenario that the individuals tend to discount the larger-but-delayed reinforce relative to the smaller-but-immediate one when they are required to choose between such alternatives(Green & Myerson,2004).A robust body of empirical researches suggested that the delay discounting are prone to be observed in humanity and even in quadrumanas.A wide range of evidences indicated that the tendency of temporal discounting can potently predict the long-term academic achievement,health,fertility,education,marriage,investment,subjective well-being and objective socio-economic status,and can further bias decisions of public policy,strategic decision,environmental protection,and global warming(Heil,Johnson,Higgins,& Bickel,2006;Bickel et al.,2007;Weller,Cook,Avsar,& Cox,2008).Obviously,it is potently worthwhile to improve the quality of our life and provide the guidance for national decision-makings and strategies in terms of the in-depth investigation and exploration.More importantly,the substantial development of non-invasive neuroimaging technique and the full-blown application of functional magnetic resonance imaging(f MRI)facilitate us to further uncover the nature and processes of delay discounting with respect to the neuronal understructure.Of note,over the course of the past several decades,even though the researches concerning the neural subtracts of delay discounting have proliferated and obtain comparatively fruitful evidences and findings,the understanding of delay discounting based on the human connectome and graph-theoretical algorithms is still lacking.Sporns and his colleagues have proposed that the neuronal mechanisms of cognitive behaviors can be recognized into three hierarchical profiles:micro-scale domain(i.e.,neuron and neurotransmitter),mecso-scale domain(i.e.,cortical column)and macro-scale domain(i.e.,brain regions).With specific regard to macro-scale profile,the current studies have explored the location of brain regions that involve in delay discounting and even attest the relevant large-scale brain sub-networks for devalued behaviors.However,the topological pattern of these neural networks relating to delay discounting based on the human connectome and complex graph-theoretical algorithms is not accounted adequately.Thus,to fill this gap,we attempt to estimate the dynamics of “scale-free” property of nine brain intrinsic sub-networks(ICs)of resting-state and calculate the overall organization(i.e.,small-worldness)of whole-brain functional connectivity networks for the sake of the better understanding on the association between the topological metrics and delay discounting.Specifically,in the study 1,we undertook the independent component analysis(ICA)to decompose the blind signals derived from time-series of resting-state f MRI scans in one sample with large sample size(105 participants)for extracting nine ICs.Then the“scale-free” property of these networks was quantificationally measured by the Hurst exponent(H),with higher H for stronger “scale-free” property.Finally,the partial correlation analysis was conducted to exam the association between H and discounting rates(measured by the area under curve).In the study 2,we utilized the multi-scale atlas(i.e.,AAL-90 and Power-264)to re-construct the whole-brain networks and calculate them with the “small-worldness” based on the graph-theoretical algorithm.Afterward,the partial correlation model was also conducted for investigating the association between discounting rates and this topological metric.Overall,these studies performed the graph-theoretical analysis to identify the topological properties of brain networks(i.e.,scale-free and small-worldness)in favor to the understanding of neuronal subtract of delay discounting.In the study 1,to unravel the neuronal mechanism of delay discounting in terms of “scale-free”property,we conducted the ICA to extract nine ICs from the blind signals of resting-state f MRI scans,and calculated the Hurst exponent(H)by Wavelet Vanishing Moment(WVM)transformation to represent the magnitude of this attribute.Finally,the partial correlation model was built to estimate the association between H and delay discounting rates after control of personality,ages,education and etc.The outcomes revealed the significantly negative correlation between AUC and H of default mode network(DMN,r(105)=-0.487;Bonferroni-correct,p<0.05),whilst we also identified the parallel results in the salience network(SAN).In other words,the higher “scale-free” organized pattern is comparable to steep discounting rates.In the study 2,to re-construct the whole-brain networks,the multi-scale atlas(i.e.,AAL-90 and Power-264)were conducted as nodes for the graph(network).Subsequently,we further calculated the Pearson correlational coefficient between each pair of nodes as the conditions to determine whether the edges existed in this graph.This procedure preserved the outcomes from the potential distortion due to the asymmetry of network resolution.In addition,two independent samples were recruited for the sake of the cross-validation of these findings,including 53 participants for principal sample and 52 participants for the replicated sample.Afterward,the graph-theoretical analysis and Markov algorithm was performed to evaluate the clustering coefficient and shortest averaged path for characterize the organization of “small-world” in these generated whole-brain networks.Finally,the partial correlation model was conducted to investigate the association between“small-worldness” and delay discounting rates.Our results have demonstrated as below:(1)the high reproducibility of network constructions with AAL-90 atlas(r = 0.921,p < 1×10-5,95% CI:0.917-0.925)and Power-264 atlas(r = 0.938,p < 1×10-5,95% CI: 0.937-0.939)was clarified;(2)the“small-worldness” was identified in these generated whole-brain networks over the range of 6 % to40 % of the sparsity,irrespective of which atlas(i.e.,AAL-90 or Power-264)was conduct(γ > 1,λ ≈1,σ > 1);(3)the positive correlation between “small-worldness”(σ)and delay discounting rates was observed in whole-brain network;(4)the significantly positive correlation between local efficiency of whole-brain networks and delay discounting rates was found with both AAL-90 and Power-264 atlas,but this association was absent in the global efficiency.In summary,two main conclusions can be denoted here:(1)the “scale-free” dynamics of large-scale intrinsic brain sub-networks,namely DMN and SN,was positively correlated with delay discounting rates,with higher scale-free property for steep underweighted tendency on the delayed reinforce;(2)the “small-worldness” of whole-brain connectivity networks was negatively correlated with discounting rates,with stronger “small-worldness” communicative pattern for less devaluation.Together,our studies uncovered the topologically neuronal pattern of delay discounting in terms of“scale-free” dynamics of DMN/SN and the “small-world” organization of whole-brain connectivity networks.Given these novel evidences based on the complex graph-theoretical algorithms,our findings extended the understanding of the topologically neuronal pattern of delay discounting,thus providing the opportunity to shed light on the natures and theories of discounting behavior.