Alterations of Cerebral Hemodynamics and Network Properties Induced by Newsvendor Problem in the Human Prefrontal Cortex

Hashini Wanniarachchi; Yan Lang; Xinlong Wang; Tyrell Pruitt; Sridhar Nerur; Kay-Yut Chen; Hanli Liu

doi:10.3389/fnhum.2020.598502

Alterations of Cerebral Hemodynamics and Network Properties Induced by Newsvendor Problem in the Human Prefrontal Cortex

Front Hum Neurosci. 2021 Jan 15:14:598502. doi: 10.3389/fnhum.2020.598502. eCollection 2020.

Authors

Hashini Wanniarachchi¹, Yan Lang², Xinlong Wang¹, Tyrell Pruitt¹, Sridhar Nerur², Kay-Yut Chen², Hanli Liu¹

Affiliations

¹ Department of Bioengineering, University of Texas at Arlington, Arlington, TX, United States.
² Department of Information Systems and Operations Management, University of Texas at Arlington, Arlington, TX, United States.

Abstract

While many publications have reported brain hemodynamic responses to decision-making under various conditions of risk, no inventory management scenarios, such as the newsvendor problem (NP), have been investigated in conjunction with neuroimaging. In this study, we hypothesized (I) that NP stimulates the dorsolateral prefrontal cortex (DLPFC) and the orbitofrontal cortex (OFC) joined with frontal polar area (FPA) significantly in the human brain, and (II) that local brain network properties are increased when a person transits from rest to the NP decision-making phase. A 77-channel functional near infrared spectroscopy (fNIRS) system with wide field-of-view (FOV) was employed to measure frontal cerebral hemodynamics in response to NP in 27 healthy human subjects. NP-induced changes in oxy-hemoglobin concentration, Δ[HbO], were investigated using a general linear model (GLM) and graph theory analysis (GTA). Significant activation induced by NP was shown in both DLPFC and OFC+FPA across all subjects. Specifically, higher risk NP with low-profit margins (LM) activated left-DLPFC but deactivated right-DLPFC in 14 subjects, while lower risk NP with high-profit margins (HM) stimulated both DLPFC and OFC+FPA in 13 subjects. The local efficiency, clustering coefficient, and path length of the network metrics were significantly enhanced under NP decision making. In summary, multi-channel fNIRS enabled us to identify DLPFC and OFC+FPA as key cortical regions of brain activations when subjects were making inventory-management risk decisions. We demonstrated that challenging NP resulted in the deactivation within right-DLPFC due to higher levels of stress. Also, local brain network properties were increased when a person transitioned from the rest phase to the NP decision-making phase.

Keywords: brain network; functional near infrared spectroscopy; graph theory; newsvendor problem; risk decision-making.