Evidence for criterion shifts in visual perceptual learning: data and implications.

Area of Science:

• Visual perception
• Cognitive psychology
• Neuroscience

Background:

• Visual perceptual learning (VPL) for contrast detection is known to decrease detection thresholds.
• These threshold decreases are consistently accompanied by increases in false alarm rates.
• The robustness and underlying mechanisms of these VPL-induced changes require further investigation.

Purpose of the Study:

• To assess the robustness and replicability of VPL-induced changes in detection thresholds and false alarm rates.
• To determine if these changes are independent of specific task demands and feedback conditions.
• To elucidate the sources of increased false alarm rates, exploring shifts in sensitivity and bias.

Main Methods:

• Participants underwent perceptual practice and threshold estimation using varied methods.
• The study manipulated the presence or absence of trial-by-trial feedback.
• Data were analyzed to differentiate changes in sensitivity (d') and response bias (criterion).

Main Results:

• VPL-induced increases in false alarm rates were robust and replicated across different task demands and feedback conditions.
• The source of increased false alarms was attributed to shifts in both sensitivity and bias.
• Alternative explanations not requiring strategic shifts in response criteria were also considered.

Conclusions:

• Visual perceptual learning for contrast detection is a robust phenomenon.
• Learning-related increases in false alarms are not solely due to strategic response criterion shifts.
• VPL may involve changes beyond perceptual sensitivity, potentially engaging cortical circuits beyond primary visual areas.