The neurophysiological basis of leadership: a machine learning approach

Abstract

PurposeThis research employed two neurophysiological techniques (electroencephalograms (EEG) and galvanic skin response (GSR)) and machine learning algorithms to capture and analyze relationship-oriented leadership (ROL) and task-oriented leadership (TOL). By grounding the study in the theoretical perspectives of transformational leadership and embodied leadership, the study draws connections to the human body's role in activating ROL and TOL styles.Design/methodology/approachEEG and GSR signals were recorded during resting state and event-related brain activity for 52 study participants. Both leadership styles were assessed independently using a standard questionnaire, and brain activity was captured by presenting subjects with emotional stimuli.FindingsROL revealed differences in EEG baseline over the frontal lobes during emotional stimuli, but no differences were found in GSR signals. TOL style, on the other hand, did not present significant differences in either EEG or GSR responses, as no biomarkers showed differences. Hence, it was concluded that EEG measures were better at recognizing brain activity associated with ROL than TOL. EEG signals were also strongest when individuals were presented with stimuli containing positive (specifically, happy) emotional content. A subsequent machine learning model developed using EEG and GSR data to recognize high/low levels of ROL and TOL predicted ROL with 81% accuracy.Originality/valueThe current research integrates psychophysiological techniques like EEG with machine learning to capture and analyze study variables. In doing so, the study addresses biases associated with self-reported surveys that are conventionally used in management research. This rigorous and interdisciplinary research advances leadership literature by striking a balance between neurological data and the theoretical underpinnings of transformational and embodied leadership.