An Episodic Model of Task Switching Effects: Erasing the Homunculus from Memory

Area of Science:

• Cognitive psychology
• Computational neuroscience
• Artificial intelligence

Background:

• Human performance in speeded response time tasks is complex.
• Existing models struggle to capture nuanced cognitive behaviors.
• The Parallel Episodic Processing (PEP) model offers an exemplar-based learning approach.

Purpose of the Study:

• To extend the PEP model to incorporate instruction following (task rules and goals).
• To simulate key findings in the domain of task switching.
• To evaluate the PEP model's explanatory power against competing models.

Main Methods:

• Designed an extended Parallel Episodic Processing (PEP) model capable of following instructions.
• Simulated human performance data from various task switching experiments.
• Compared PEP model simulations against empirical data and alternative models.

Main Results:

• The extended PEP model accurately simulated switch costs, congruency effects, and repetition asymmetries.
• The model captured cue repetition benefits and complex patterns in cued task switching.
• PEP demonstrated superior fit to data compared to several competing task switching models.

Conclusions:

• The PEP model provides a parsimonious and powerful explanation for task switching phenomena.
• Episodic accounts, particularly feature-integration biases, are crucial for understanding task switching.
• The PEP model's success challenges previous assumptions about the primary drivers of task switching performance.