What a late-shift decline actually looks like
In a large urban court system, judges' release rates declined modestly in the hours before lunch and before dinner, and the pattern held after adjusting for many observable case and defendant characteristics [1]. That is a concrete, measurable shift in a high-stakes decision: as a session wore on, judges were somewhat less likely to release defendants. The size was modest, not dramatic.
In hospital and emergency settings, the picture is murkier. A scoping review of decision fatigue in hospitals notes that end-of-shift timing may affect how many patients are seen, how acute they are, and what investigations are ordered [3]. But a narrative review of emergency physicians reports that while the number and complexity of decisions correlate with self-reported fatigue, the effect on clinical decision quality is uncertain, and a subset of studies found no clear relationship between decision fatigue and errors in clinical judgment [4].
The meal-break test: a natural experiment for 'running out'
If late-shift decline were simply mental fuel running low, a break with food and rest should refill the tank and restore decision quality. The pretrial release study tested exactly this and found that release rates remained unchanged after a meal break, even though judges had the chance to rest and eat [1]. The decline before lunch and dinner also did not continue through the rest of each session—arraignment time had no effect in the remainder of a decision-making session [1].
That pattern is hard to square with a simple depletion story. The authors conclude that to the extent decision fatigue plays a role in these judgments, the effects are small and inconsistent with psychological depletion processes [1]. In other words, the most direct test of 'resources run out' did not behave the way that theory predicts.
A metabolic explanation: cost goes up, not fuel goes to zero
A different line of evidence points to a metabolic mechanism rather than a simple empty tank. In a study using magnetic resonance spectroscopy to track brain chemistry across an approximate workday, participants who did high-demand cognitive control tasks showed a shift toward short-delay and little-effort options, along with reduced pupil dilation during decisions—a low-cost bias captured with computational modeling [5]. At the end of the day, the high-demand group had higher glutamate concentration and glutamate/glutamine diffusion in the lateral prefrontal cortex, a cognitive control brain region, compared with the low-demand group and a reference brain region [5].
The authors propose a neuro-metabolic account: glutamate accumulation may trigger a regulation mechanism that makes prefrontal activation more costly, which would explain why cognitive control is harder to mobilize after a strenuous workday [5]. This is a proposed mechanism, not a settled cause. It also reframes the question: the issue may not be that resources are gone, but that exerting control has become more expensive, which could push people toward cheaper, more automatic choices.
Is it shortcuts, depletion, or something else?
The 'mental shortcuts' explanation is plausible but not directly established by these studies. The pretrial release study found that judges' rate of concurring with prosecutorial bail requests did not appear to be influenced by arraignment time or a meal break [1]—a result that does not fit neatly with either a depletion or a simple shortcut story. The UAV operator study reports a significant decline in decision quality after six hours of continuous work, with an average score of 68.2 at the end of an eight-hour shift, but it does not address whether the decline comes from resource depletion or from mental shortcuts [2].
Reviews in hospital and emergency medicine reach a similar impasse: decision quality assessments exist, but explicit evidence on depletion versus shortcuts as causes is lacking [3][4]. The emergency medicine review notes that some researchers argue decision fatigue may be mitigated by experience, training, and support systems [4], which would suggest the effect is not a fixed biological limit. Taken together, the evidence supports a modest, context-dependent decline in some settings, but the mechanism—running out versus switching to shortcuts—remains unresolved, and the meal-break result argues against the simplest version of the depletion story [1].
Sources used in this answer
Pretrial release judgments and decision fatigue
In a large urban court system, judges' pretrial release rates declined modestly before lunch and dinner, but release rates were unchanged after a meal break and arraignment time had no effect in the remainder of each session, which the authors say is inconsistent with psychological depletion explanations.
The impact of cognitive load on tactical decision-making of unmanned aerial vehicle operators during extended shifts
UAV operators showed a significant decline in tactical decision quality after six hours of continuous work, with an average score of 68.2 at the end of an eight-hour shift, but the study did not test whether this was due to resource depletion or mental shortcuts.
Decision fatigue in hospital settings: A scoping review
A scoping review of decision fatigue in hospital settings reports that end-of-shift timing may affect patient acuity, patient numbers, and investigations ordered, but does not provide explicit evidence on depletion versus shortcuts as causes.
Decision Fatigue in Emergency Medicine: An Exploration of Its Validity
A narrative review of emergency medicine finds that decision complexity and volume correlate with self-reported fatigue, but the effect on clinical decision quality is uncertain, with some studies finding no clear relationship between decision fatigue and errors.
A neuro-metabolic account of why daylong cognitive work alters the control of economic decisions
After a day of high-demand cognitive control tasks, participants shifted toward short-delay and low-effort economic choices and showed higher glutamate concentration in the lateral prefrontal cortex, supporting a proposed neuro-metabolic account in which accumulated glutamate makes cognitive control more costly.