To characterize contrast detection, noise should be extended, not localized

Area of Science:

• Visual Neuroscience
• Perceptual Psychology

Background:

• Adding noise is crucial for characterizing visual processing.
• Different noise types (extended vs. 0D) have been proposed to avoid processing strategy shifts or cross-channel suppression.
• Internal noise in visual systems is generally considered extended across dimensions.

Purpose of the Study:

• To argue that extended noise does not contaminate contrast thresholds via cross-channel suppression.
• To review evidence that 0D noise employs different processing strategies compared to no noise or extended noise.
• To recommend the use of extended external noise for visual processing studies.

Main Methods:

• Theoretical argument regarding contrast gain in extended noise.
• Review of empirical findings on target detection in different noise conditions (0D vs. extended).
• Comparison of processing strategies in the presence of various external noise types and internal noise.

Main Results:

• Contrast thresholds in extended noise are not affected by noise-induced cross-channel suppression.
• Contrast gains proportionally affect both signal and noise in extended noise, preserving the signal-to-noise ratio.
• Target detection in 0D noise utilizes distinct processing strategies compared to detection in quiet or extended noise.

Conclusions:

• Extended noise is suitable for studying visual processing without introducing confounds like cross-channel suppression.
• The processing strategy for target detection differs significantly between 0D noise and other conditions.
• Consistent with the nature of internal noise, extended external noise is recommended for visual studies assuming consistent processing strategies.