Decision criteria in dual discrimination tasks estimated using external-noise methods

Area of Science:

• Psychology
• Cognitive Science
• Neuroscience

Background:

• Classical signal detection theory (SDT) relies on internal noise, limiting unambiguous measurement of decision criteria.
• Existing experimental paradigms struggle to isolate decision criteria from response bias and internal noise.

Purpose of the Study:

• To introduce and validate a novel method using external noise to estimate decision criteria and their variability in stimulus space.
• To investigate criterion interactions and observer strategies in single and dual tasks.
• To test hypotheses regarding criterion collapse versus equal false alarm rates in dual-task scenarios.

Main Methods:

• Employed a unidimensional external noise paradigm to define and measure decision criteria in stimulus space.
• Assessed criterion variability relative to sensory thresholds.
• Compared observer performance in single-task versus dual-task conditions.
• Investigated the role of false alarm (FA) rates in criterion setting.

Main Results:

• The external noise method successfully estimated criteria and their variability, with criterion variability approximately 1.6 times the sensory threshold.
• In dual tasks, observer criteria exhibited an "attraction" effect, deviating from predictions for independent criteria.
• Observers actively maintained distinct FA rates, even when SDT predicted equal rates.
• Psychophysical training improved criterion separation in dual tasks.

Conclusions:

• External noise provides an unambiguous measure of decision criteria, overcoming limitations of internal noise-based methods.
• Observer criteria interact in dual tasks, showing an attraction effect not fully explained by simply equating FA rates.
• SDT predictions for dual tasks require refinement to account for observers' strategies in maintaining distinct FA rates and criterion separation.