Logistic resource limitation model for quasi real-time measured subjective cognitive load predicts Hill function of hemoglobin-oxygen saturation

Abstract

Cognitive processing and memory resources invested in task execution determine mental workload (MWL) that is quantified through objective physiological measures such as heart rate and variability, EEG, and hemoglobin oxygen (HbO2) saturation, and subjective methods like periodic quasi-real-time "instantaneous self-assessment" (ISA) with discrete five- or seven-level WL-scales. Previously published results of human-in-the-loop (HITL) air-traffic control simulations with highly trained domain experts provided initial evidence for logistic and power law functional dependencies between subjective MWL self-assessment reports and simultaneously monitored task load and simulation variables (e.g. communication and traffic load). Here we show that a biased "Logistic Resource Limitation" (LRL) model for regression based parameter estimates of subjective self-reports through combination with a logistic task load function leads to a cognitive power law with parametric correspondence to the classical Hill function that quantifies HbO2 saturation. Hill function saturation exponent and equilibrium dissociation constant turned out to show surprising agreement with corresponding estimates of the power law parameters derived from the LRL-model applied to published independent data sets from the three different HITL-simulation experiments. Our results suggest the hypothesis that under certain conditions quasi real-time subjective (behavioral) reporting of cognitive load due to task execution might represent the output of an interoceptive HbO2 saturation sensor that measures resource limitation of neural energy supply. From the HbO2 - saturation perspective, our results might provide an additional aspect to the "selfish brain" theory for cortical energy supply as derived by A. Peters et al. based on a logistic Glucose push-pull supply chain model. However, more focused experiments are required including direct (e.g. fNIR based) measurements of HbO2-saturation to further support (or falsify) our conclusions. Author SummaryMeasurements of mental workload of domain experts under cognitive task requirements by human-in-the-loop simulation experiments utilize subjective and objective methods and measures. Standard data analysis is mostly limited to linear statistical methods such as variance and regression analysis for quantifying load differences under different task requirements. Based on nonlinear resource limitation models with asymptotic saturation limits we derive here a cognitive power law for the dependency of real-time subjective work- vs. objective task load. The focused analysis of three previously published independent datasets revealed an unexpected formal and quantitative equivalence with the classical Hill-function of blood-oxygen saturation. Our results suggest the hypothesis of a close quantitative relationship between subjective load reports and an interoceptive senor for cortical energy resources.