Sequence-dependent sensitivity explains the accuracy of decisions when cues are separated with a gap

Area of Science:

• Cognitive neuroscience
• Computational neuroscience
• Decision science

Background:

• Perceptual decision-making involves integrating information from stimuli presented with varying temporal gaps.
• Existing models fail to explain human performance in tasks involving two discrete pulses, which exceeds predictions of perfect accumulators.
• The neural mechanisms for integrating information across temporal gaps remain unclear.

Purpose of the Study:

• To investigate the neural mechanisms underlying information integration in decision-making tasks with temporal gaps.
• To explain the superior performance observed in humans when integrating information from two discrete pulses.
• To identify a computational model that accurately captures human behavior in these tasks.

Main Methods:

• Utilized a set of modified drift-diffusion models (DDMs) based on different hypotheses.
• Employed model comparison techniques to evaluate the explanatory power of each DDM.
• Focused on analyzing the impact of temporal gaps and pulse order on information accumulation.

Main Results:

• Demonstrated that accumulated information from earlier evidence dynamically affects the accumulation of later evidence.
• Found that the rate of information extraction is dependent on whether a pulse is the first or second in a sequence.
• Showed that a DDM with a dynamic drift rate can account for the enhanced effect of the second pulse on decisions.

Conclusions:

• The integration of evidence in decision-making is not a simple accumulation process but is influenced by the temporal sequence of information.
• A dynamic drift rate in DDMs is crucial for accurately modeling human decision-making, particularly the influence of sequential evidence.
• Findings provide a more refined understanding of the neural computations supporting perceptual decisions under temporal uncertainty.