Neural oscillation as a selective modulatory mechanism on decision confidence, speed and accuracy

Abstract

Neural oscillations have been associated with decision-making processes, but their underlying network mechanisms remain unclear. This study investigates how neural oscillations influence decision network models of competing cortical columns with varying intrinsic and emergent timescales. Our findings reveal that decision networks with faster excitatory than inhibitory synapses are more susceptible to oscillatory modulations. Higher in-phase oscillation amplitude reduces decision confidence without affecting accuracy, while decision speed increases. In contrast, anti-phase modulation increases decision accuracy, confidence and speed. Increasing oscillation frequency reverses these effects. Changing oscillatory phase difference gradually modulates decision behaviour, with decision confidence affected nonlinearly. Moreover, neural resonance can further enhance modulatory susceptibility for network with faster excitatory than inhibitory synapses. These effects decouple decision accuracy, speed and confidence, challenging standard speed-accuracy trade-off. These phenomena can be explained by excitatory neural populations contributing more to in-phase modulation, while inhibitory neural populations to anti-phase modulation. State-space trajectories momentum swinging with respect to network steady states and decision uncertainty manifold further provide insights into the neural circuit mechanisms. Our work provides mechanistic insights into how neurobiological diversity shapes decision-making processes in the presence of ubiquitous neural oscillations. Significance StatementNeural oscillations shape how the brain balances decision speed, accuracy, and confidence. Here, we show that tuning oscillatory amplitude, frequency, or phase can selectively alter these decision measures. In some cases, these modulations even break the usual trade-off between speed and accuracy, revealing a more flexible decision-making mechanism than previously assumed. Our findings highlight how synaptic time scales and rhythmic brain activity can give rise to distinct patterns of decision performance. This insight may guide new strategies for improving decisional processes in both healthy populations and clinical conditions.