The Neural Oscillatory Mechanisms of Emotion Affecting Working Memory: a Dual-path Model Based on Competition and Interference

Working memory (WM) serves as the core of advanced cognitive functions, enabling the temporary storage and manipulation of information, which is crucial for reasoning, comprehension, and decision-making. However, its performance is influenced by emotional states; for instance, stress or anxiety may impair accurate recall or prioritize the processing of threatening stimuli. This review integrates research on the neural oscillatory mechanisms by which emotions affect WM, emphasizing shared patterns in the θ, α, β, and γ frequency bands. Emotions activate distinct neural circuits and alter oscillatory characteristics through arousal and valence, overlapping with the neural activities required for WM processes. This article aims to elucidate these mechanisms and propose a dual-pathway theoretical framework for emotional influences on WM. The cognitive efficiency hypothesis posits that emotions and WM compete for shared oscillatory resources. Both emotional processing and high WM load enhance cortical excitability by reducing α power to optimize attention allocation. Negative emotions, under sufficient presentation time, prioritize resource allocation to improve WM precision through enhanced sustained α suppression, albeit at the cost of reducing the number of remembered items. The patterns of θ power increases induced by emotions and task load overlap, potentially leading to θ/β saturation under high load, which limits cognitive regulation, while under low load, they synergistically support information maintenance. Emotional states and the β oscillations relied upon for WM maintenance converge in frequency, prone to synergy or competition. High arousal induced by emotions, as well as anxiety alleviation through music, can reduce prefrontal β oscillation power, bringing the brain closer to the low‑β state required for WM maintenance, thereby enhancing memory performance and neural efficiency. Similar γ oscillations and α‑γ coupling patterns induced by emotions and WM compete for neural resources, interfering with inter-brain region information integration and weakening the regulation of γ amplitude by α rhythms, thus impairing information exchange efficiency and processing stability. The interference hypothesis suggests that emotions directly disrupt the rhythms required for WM. Negative emotions typically reduce the power and synchrony of α oscillations, a change opposite to the α enhancement mode needed during the WM maintenance phase, thereby weakening interference shielding capabilities. The decrease in θ oscillation power induced by negative emotions contrasts with the θ enhancement required for WM tasks, thereby interfering with information maintenance, manipulation, and multi-item integration, leading to declines in memory capacity and precision. Negative or high-arousal emotions significantly disrupt β oscillation power and synchrony in WM tasks, resulting in impaired inter-brain coordination, reduced information stability, and consequently weakened task performance. Emotions directly alter γ oscillation power and θ-γ coupling patterns, with these changes opposing the enhancement direction required for WM tasks, thereby causing neural activity imbalances and information integration obstacles. The cognitive efficiency hypothesis and the interference hypothesis, as a dual-pathway model for emotional influences on WM, are not opposing explanatory frameworks but rather reflect a dynamic and complementary regulatory mechanism in emotion-cognition interactions. The brain flexibly selects interference or efficiency pathways based on current task load, emotional intensity, and individual states to maintain overall functional stability, with its core mechanism lying in the limited and adaptive allocation of brain neural oscillatory resources.Future research should delve into the impacts of emotional states, cognitive load, emotional arousal, and regulatory strategies on WM and its neural modulatory effects, to optimize personalized cognitive intervention strategies.