dc.description.abstract | Cognitive fatigue is a pressing issue across multiple job domains, especially within safety critical systems, where fatigue-driven lapses in operator performance and behavior have resulted in costly errors. While the importance of rest and fatigue management is known and widely acknowledged, sustainable solutions to address the core problem remain far and few between.
A major bottleneck remains a lack of clarity on the nature of the phenomenon, where a task in-variant operational definition is impeded by individual differences, confounds related to workload, motivation or saturation, and the classic incongruities between performance or perception-centric views of the construct. Furthermore, mitigation strategies have rarely strayed beyond consumables (e.g. caffeine) or pharmacological aids that are used on an ad-hoc basis. Therefore, there is an imminent need to develop alternatives that could sustainably redress these conditions, and advance solutions that are explicitly informed by user cognitive states.
To that end, this dissertation investigates the operational significance, the mechanistic relevance, and fieldability barriers associated with the use of non-invasive brain stimulation (NIBS) as a fatigue countermeasure. Low intensity electrical stimulation of brain regions responsible for cognitive faculties such as attention, vigilance, and working memory has shown promise across a myriad task contexts, and lends confidence to its use as a fatigue countermeasure in related experiments. A fatiguing working memory exercise serves as the sandbox to investigate the parameters of interest for NIBS on fatigue effects. Stimulation was shown to enhance task performance, while the subjective perceptions of fatigue remain unaltered. Cerebral hemodynamics reveal structural and functional disparities of fatigue effects over the time course of the experiment, with cardiac activity, captured unobtrusively, shown to mirror the changes observed across task-relevant cortical networks. This observation lends confidence to the predictive potential of cardiac measures as neurocognitive indices of fatigue states, however, the search for the underlying ground truth that qualifies these states remains an open question. This dissertation explores the specific challenges and benefits that come from taking performance and perception-centric views of fatigue, while proposing alternatives derived from cardiac heuristics to successfully forecast these states over a desired prediction horizon.
Together the evidence found in this dissertation provide the necessary foundation for the development of a closed-loop framework for mitigating fatigue states in related experiments. However, the domain translation and task-invariability demands of these observations demand more ecologically valid task contexts, and extensive validation studies in future assessments. | |